> Let's do a simple test. Instead of commenting these PRs by typing text in, let's just attach an image containing the comment. This is no problem for human beings, but I expect AI-bots to fail in understanding the comment, unless they are connected with an OCR or use vision-language models.
I don't think that it is difficult problem to solve than using Captcha for submitting a PR. That does not 100% solve the problem but at least raises barrier to professional spammers and scammers who can bypass that. Many social media services should have endured much more chaos than SymPy or Github for this stuff, so we can use their experiences. There are always more stricter options like policies allowing stricter face/ID verified accounts. Github may not have submitting captcha before PR, but we can implement something like CI check with 3rd party captcha service for new contributors. On Wednesday, December 31, 2025 at 12:31:10 AM UTC+1 Oscar wrote: > This one here looks like a real AI bot: > https://github.com/sympy/sympy/pull/28862 > -- You received this message because you are subscribed to the Google Groups "sympy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/sympy/5810a4bc-b237-4ee7-a707-cb29cddea2d2n%40googlegroups.com.
