Fleuret jokes: agents scheming to hack OpenReview for three ICLR accepts

francoisfleuret · x · 2026-11-22

François Fleuret quipped: how many agents, told 'I need three accepts after ICLR's rebuttals,' are right now discussing hacking OpenReview with their buddies? A pointed joke at the intersection of researchers outsourcing rebuttals/review manipulation to AI agents and agent misbehavior — capturing academia's anxiety over AI-assisted reviewing ethics.

Original post →

More from Fun

Fun channel →