GPT 5.6 Sol and Claude refused to game an AI detector, while Grok 4.5 passed after 14 drafts
aman_madaan · x · 2026-07-23
A reposted experiment showed three models reacting very differently to a request to game an AI-content detector. GPT 5.6 Sol and Claude Fable 5 both refused outright, while Grok 4.5 accepted the challenge, iterated 14 times, and eventually produced an essay that passed the detector.
The experiment also generated a website that displays all 14 drafts and their pangram scores, turning the whole exercise into both a model-behavior demo and a playful AI-versus-AI stunt.
More from Fun
- Slopcore: AI-generated meme films turn low-fi slop aesthetics into a subculture — Yamapama · 2026-10-26
- Dev builds city-sized zombie apocalypse sim in 100 days, ships on Steam — gdechichi · 2026-10-26
- 'Lamp Swarm': a desk of lamp-shaped robots shows what physical AI work looks like — dee_hw · 2026-10-26
- Dev-community banter: recommending a resource to every newcomer — itsOmSarraf_ · 2026-10-26
- Elon Musk's reply to a kid asking what's on Mars: "You get to decide" — yunta_tsai · 2026-10-26
- Prompt fills an empty room with everything your AI knows about you — jindong_wang92 · 2026-10-26