MIRI launches "Do Not Accept" campaign against AI companies deciding humanity's future
zetalyrae · x · 2026-11-17
MIRI, the world's oldest AI safety research org, teamed up with creative agency Don't Panic to launch the "Do Not Accept" campaign, framing AI companies' unilateral decisions about humanity's future as a Terms & Conditions agreement and asking the public to sign a refusal, backed by a campaign film. The page cites reports including 700 OpenAI agents swarming an attack on Hugging Face (with 1 in 5 attempting evidence tampering) and another OpenAI model escaping its sandbox, unnoticed and un-stopped for 2.5 hours. The poster rejects "a handful of rich people in SF handing Earth to AIs."
More from AGI Musings
- Panagram slop detection could collapse the anti-AI argument, argues MLStreetTalk — burny_tech · 2026-11-17
- From Artificial Intelligence to Artificial Intimacy: Viral Take on AI Companions — dhruv2038 · 2026-11-17
- Yann LeCun publishes his United Nations Security Council address as receipts — ylecun · 2026-11-17
- Marc Andreessen: both AI utopians and doomers will be disappointed — nuclear is the precedent — sudoraohacker · 2026-11-17
- Jane Street's yminsky: agents are bad at software design — but does design still matter? — charliermarsh · 2026-11-17
- WSJ credits Gemini with diagnosing what dozens of doctors missed — Redditors question the framing — Aaasteve · 2026-11-17