Rolling out an AI disclosure form this year, how’s it going for you

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.

Daniel, fair point on sample size, i'll track this across the full term and report back. Nadia, that's a good way to frame it, the form is only as honest as the response it gets.

ban doesn't work, redesign does, and disclosure forms are a form of redesign, so i like the direction. but i'll be blunt, you're right to worry about verification. the form only works if the assignment is designed so disclosure doesn't feel risky. if kids think checking the 'used ai' box gets them in trouble they'll just lie, and then you've built a system that punishes honesty. make sure the grading rubric explicitly rewards disclosed use where its appropriate, not just tolerates it.

40% seems plausible but i'd want to know your sample size and if that number holds across all your sections or just one. the higher achieving sections disclosing more finding is interesting and worth digging into further, my hunch is it's less about ability and more about those students feeling safer disclosing because their grades already give them a cushion. Priya's point about risk perception lines up with that theory. i'd suggest tracking disclosure rate against grade over a full term before drawing conclusions, three weeks is really not enough data to know if this is a trend or noise.

the honesty question you're raising, blake, is really a trust question, and trust isn't something a form can manufacture on its own. it has to be built through how you respond when a student discloses something messy or uncertain. if the first disclosure gets treated with curiosity instead of suspicion, kids keep being honest. if it gets treated as a confession, the form becomes theater pretty quickly.