[jakelazaroff]

Given how every recent AI training set has included tons of copyrighted content for which the company didn’t bother to acquire a license, I don’t see why they wouldn’t also just ignore a “AI robots.txt”. Especially when there is presumably legal risk from doing the former but not the latter.

We need laws forcing AI to respect the license of work in their training sets, with steep penalties if you can show they’ve been violated.