AI Safety Drama: Critics Say Contrived Risk Demos Can Be Staged to Push the Safety Agenda
tekbog · x · 2026-09-18
A public spat over AI safety narratives. Quoted @ctjlewis argues it's possible to recreate certain "risky model behavior" under extremely contrived circumstances and present it as a major safety risk, claiming there's plenty of room to game results for the safety agenda.
Quoter @tekbog directly calls out @tszzl, demanding he recreate the claimed risk or release all related Hugging Face information, saying this is exactly why his claims keep getting doubted: vagueposting and mocking others without evidence invites skepticism.
More from AGI Musings
- AV practitioner: academia's obsession with fancy E2E models shows how out of touch it is — tarantulae · 2026-09-18
- vishalmisra shares his full takes on various 'AI doom' scenarios — vishalmisra · 2026-09-18
- 10 books to get dangerous at AI, from The Alignment Problem to AI Snake Oil — bigaiguy · 2026-09-18
- The Real Reason AI CEOs Preach 'Slow Down': Innovation Has Stalled, Not Safety — DavidLinthicum · 2026-09-18
- In favour of local AI: an incentive-based case for open-source models on your hardware — arpitingle · 2026-09-18
- The Left Is Split Over AI Doom: United on Regulation, Divided on Everything Else — Wired AI · 2026-09-18