Hi, As those who were at the meeting two weeks ago are aware, there has been some ongoing discussion about the use of AI in code reviews (first on https://review.coreboot.org/93682, then in the meeting). I have now uploaded a draft for new Gerrit guidelines that put some rules on if and how AI can be used in the review process. The TL;DR is:
* You can use AI to post comments about new issues during review, but not to respond to others' comments * You should tag AI comments visibly (e.g. prefix with `[AI-generated]`) * You must manually review all AI comments before you submit them and make sure they make sense I'm also adding an AGENTS.md file that will hopefully get the AI to self-police some of this, although ultimately it's always the responsibility of the human to keep their AI in check. Let me know what you think and if you have any concerns or suggestions on: https://review.coreboot.org/c/coreboot/+/95178 _______________________________________________ coreboot mailing list -- [email protected] To unsubscribe send an email to [email protected]
