Reviewers should verify correctness, not AI authorship — Pangram won't get us there

ipeirotis · x · 2026-09-24

NYU professor ipeirotis argues peer review should solve for correctness first: he would gladly review only technically correct papers and focus on judging interestingness. He doesn't care whether a technically correct paper is AI-written as long as its claims are verified — and claims AI-detection tools like Pangram won't get us there.

Related event: Scholars Debate Zero-Tolerance on AI-Written Papers(10 posts)→

Original post →

More from AGI Musings

AGI Musings channel →