ChatGPT & Classroom · Posted by Angela Castillo ·

Peer Review for AI-Assisted Work – A New Approach I’m Trying

3

Drama and creative writing teacher (Ontario). Sharing something I’m piloting that might be useful for other creative writing or English teachers.

Context: I’ve accepted that students are using AI as part of their creative process. instead of trying to detect and prohibit, I’m making it explicit.

New assignment structure:
1. Students submit their draft with an AI contribution note (what AI helped with, what they wrote themselves)
2. A peer reviewer reads the draft AND the contribution note
3. Peer review specifically addresses: where does the human voice come through? where does it feel generic? what would make the student’s genuine voice stronger?

Results: better peer review conversations than I’ve had in years. students are more honest because disclosure is expected. the peer review is more specific because reviewers know what to look at. and students are thinking more carefully about their own voice vs AI voice.

it’s not solving the integrity question for all purposes but for creative writing it’s actually improving the learning.

4 replies

4 Replies

4

THIS IS BRILLIANT!! The peer review specifically looking for student voice vs AI voice is such a smart reframe!! Trying this in September. Sharing with my whole department RIGHT NOW.

6

the peer review model for AI-assisted work is underexplored. it forces students to be able to explain and defend the choices in their work, which is exactly what AI assistance can't do for them. a student who used AI to generate content they don't understand will be exposed in peer review immediately without any detection tool involved.

16

the disclosure-as-expectation design is doing something clever: it removes the incentive to hide, which means you get honest data about how students are actually working. the peer review that follows is also doing genuine metacognitive work - students have to identify what makes writing sound human vs generic. thats actually a valuable literary skill.

9

running a pilot of something similar this semester. students submit AI interaction logs alongside their drafts - showing what they asked and how they revised the output. peer reviewers evaluate both the final essay and the revision quality. early results: students are using AI more thoughtfully when they know the process is being assessed.