greCAPTCHA: testing manuscript-specific understanding to catch AI-written papers

Dr_Atoosa · x · 2026-09-17

In a thread, the author presents their paper proposing greCAPTCHA, an assessment approach that evaluates an author's manuscript-specific understanding rather than passively inspecting prose. It targets the flood of AI-assisted/AI-generated submissions hitting peer-review venues, universities, and education institutes with insufficient human oversight, where author names are no longer a reliable signal of expertise.

Related event: greCAPTCHA: Probing Manuscript Authors' Understanding with Dynamic Questions, AUC 0.90(6 posts)→

Original post →

More from Safety

Safety channel →