Bounty hits $100k as no one produces pre-2022 French writing flagged 100% AI by detectors

Afinetheorem · x · 2026-09-23

AI researcher Afinetheorem has a public standing bet on AI-text-detector reliability: produce just two pages of French literary writing provably from before 2022, by any author, that detectors score as 100% AI-generated, and win the bounty — now at $100k as his audience passed 100k followers. He says over 50,000 people have seen the bet and "not one has even tried," and argues that claims like "my 2014 thesis gets flagged as AI" or "metaphor-heavy writing triggers detection" are either lies or misunderstandings of how the technology works. The bet is effectively a public challenge to the accuracy of AI text detectors.

Related event: Bounty Raised to $100K for French Text Falsely Flagged as AI(3 posts)→

Original post →

More from Fun

Fun channel →