glibc AI policy

[Paul Eggert]

On 2026-04-16 09:27, Florian Weimer wrote:
> It really sucks if you are accused of being an LLM.  I
> don't want reviewers doing that routinely.

I don't either, but LLM usage is routine so unfortunately this issue 
comes up routinely.

I have some experience of being an LLM accuser in both academic and 
code-contribution situations, and there was never any doubt that LLMs 
were used. The accused never denied it, for what it's worth.

If accusers are being at all responsible, the vast majority of LLM 
subterfuge goes unreported, and only blatant examples get accusations. 
And as LLMs improve it becomes less obvious which uses are blatant.

In the end we are on the honor system here, just as we are with 
traditional copyright violations. The main job of reviewers in this area 
will be to remind contributors of the honor system.