[TYPES] AI-generated conference submissions

[vladi Sassone]

My 2c: I think the issue here is not moral, but practical: the moment it’ll take longer to “verify” a paper’s correctness than to “produce” it, all will break down (should be familiar to people in complexity). And it takes a long time and expertise/skill to detect abuses. (It is even worse with journal submissions.) So yes, it looks like the way forward is a “tool” to (at least) scan (or more!) all submissions pre-review. But then the “tool” will be AI too, which raises interesting questions per se… (and by the way this is the same problem all schools/universities have: it has become pointless to give programming assignments, you’ll end up spending your precious time marking work done by AI).

Cheers