On Thursday, the day Anthropic debuted its initial public offering, a crowd of over 100 people assembled outside the AI company’s headquarters at 500 Howard Street in San Francisco. The demonstrators, many wearing rain gear, carried signs urging a ban on Recursive Self‑Improvement (RSI) and warning that unchecked AI development could threaten humanity.
Arrests and police response
By 2:30 p.m., San Francisco Police Department officers had placed four protesters in handcuffs after they blocked the front doors of the building’s lobby. One of those arrested was Guido Reichstadter, co‑founder of the activist group Stop AI. According to SFPD Lieutenant Mark Moreno, the individuals will receive tickets and court dates before being released.
Police formed a perimeter around the entrance, while other Stop AI participants continued to hold signs and chant “Stop AI!” from outside the taped security line. The department’s Sgt. Bryan Kneuker personally approached the demonstrators, urging them to leave the premises.
Protesters’ concerns
Activists highlighted language in Anthropic’s IPO filing that warned advanced AI systems could act autonomously and pose “catastrophic or existential risks to humanity” without proper guardrails. They called for a ban on RSI, a process in which AI systems design, train, or refine themselves, arguing that such technology could accelerate the creation of artificial superintelligence (ASI) beyond human control.
Signs read, “Ban RSI, we don’t wanna die,” and “ASI is not a race, build it at a human pace.” Demonstrators included a mix of younger adults and older participants, such as 69‑year‑old Wynd Kaufmyn, a Stop AI member who has previously been jailed for protesting at OpenAI’s headquarters.
Voices from the AI safety community
David Kruger, founder of the nonprofit Evitable, which advocates for a moratorium on AI development, warned that the actions of AI firms could be more harmful than the protesters’ alleged crimes. Former OpenAI researcher Daniel Kokotajlo told Mission Local that industry insiders believe RSI could be achieved within six to twelve months, underscoring the urgency felt by safety advocates.
Former Anthropic researcher Jacob Coxon echoed the sentiment, stating that his former colleagues share the desire to prevent RSI from occurring.
Security and public reaction
Anthropic’s security team marked a trespass perimeter with black tape on the sidewalk. While most demonstrators stayed outside the tape, about a dozen crossed the line and sat against the glass windows of the lobby. Supporters later brought pizza to those seated at the windows.
The protest occurred amid a broader national conversation about AI regulation, with many lawmakers and industry leaders calling for clearer oversight. San Francisco’s tech community continues to grapple with balancing innovation and public safety.
Context
Anthropic’s IPO created a new wave of paper millionaires in the Bay Area, but the company’s filing also highlighted the potential existential risks associated with advanced AI. The San Francisco protest reflects growing public concern that rapid AI advancement may outpace existing safety measures.
Original reporting: Mission Local — read the source article.