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."

Original post →

More from AGI Musings

AGI Musings channel →