Over 100 protestors, including prominent AI-safety activists and "swarm chasers," gathered outside Anthropic's San Francisco headquarters, chanting and waving signs reading "Stop the AI Race" and "Don't Build AI That Builds AI."
Their complaint: Anthropic bills itself as the most safety-conscious artificial-intelligence developer but is at the same time leading what critics say is a headlong race toward so-called recursive self-improvement, sometimes called RSI, meaning AI that can improve itself.
The protest was organized by a group calling itself Ban RSI, whose website reads: "Anthropic uses AI to build better AI. They are making trillions doing so."
Jacob Coxon, the Anthropic researcher whose resignation spurred a wave of AI fears this fall, made an appearance. He called for Anthropic to lobby the government for better regulation, work with the other AI companies to coordinate on safety, and unilaterally slow down AI progress if other options fail.
"We're here because this technology is threatening our lives," Coxon said.
Comments