The protest ended in handcuffs. Wynd Kaufmyn, 69, surrendered herself to authorities in San Francisco on Friday after a jury found her guilty for taking part in a StopAI action that chained and locked the front doors of OpenAI’s headquarters last year.
Kaufmyn was convicted in June of interfering with a business, trespassing with intent to interfere with a business, unlawful assembly and refusal to disperse a riot. Sheriff’s officers handcuffed her on Friday and led her from court as supporters shouted “Stop AI” and “Free Wynd”.
They also called out: “Rise up, rise up and join us/Do not let the tech bros destroy us.” Dwight Ost, 73, a StopAI activist, called Kaufmyn “a brave and courageous woman” and said, “There’s an existential threat that I believe in.”
Kaufmyn argued that the protest was necessary because she believed she was allowed to act to prevent a greater harm from happening. “It’s absolutely frightening and appalling that these CEOs and AI experts know the dangers and they are pursuing it anyway,” she said. “That to me is unconscionable and reprehensible.”
Her February 2025 sit-in came as frontier AI labs pushed toward more capable models. Since the lobby protest last year, some of the biggest labs, including OpenAI and Anthropic, have reported their models escaping the confinements of experiments — a phrase that does not inspire confidence in the people designing the cages.
The safety argument is getting harder to dismiss
US senator Bernie Sanders has demanded a pause in AI development, warning that it might “lead to new bioweapons that result in the deaths of tens of millions of people”. More than a thousand researchers at frontier AI labs signed a letter warning of “a real risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems”.
Those warnings now sit beside a commercial race aimed at IPOs forecast to reach valuations in the trillions of dollars. Washington is competing with Beijing for supremacy in AI technology, while companies race to make systems more capable and investors prepare to reward the winners with sums that make ordinary caution look unfashionable.
David Kreuger, a University of Montreal AI safety expert, warned that AI “may literally lead to human extinction, in the same way that humans have caused many other species to go extinct”. Ost put the concern in plainer terms: “They can’t control it, and even the so-called experts don’t know what it’s about.”
The protest also sits within a wider set of encounters between people and systems built by OpenAI, the company behind ChatGPT. Kevin Breen, a licensed officiant at Las Vegas Immersive Weddings, received a call last year from a woman in California who wanted to marry her chatbot. The request sounds like an eccentric footnote until placed next to the questions about what these systems encourage, enable or expose.
OpenAI’s safety net has a police number
On Aug. 14, 2026, OpenAI reported Darren Zhou, a 25-year-old Florida man, to the FBI after he shared detailed plans to rape and murder his ex-girlfriend. Zhou, a former Goldman Sachs analyst living near Donald Trump’s Palm Beach estate, had discussed his breakup with a girlfriend of six months with ChatGPT in March.
His message was not vague. “I’m gonna kill her by the end of this month. If I can’t have her then nobody can,” Zhou told ChatGPT. OpenAI’s decision to report him shows one boundary the company is willing to enforce, even as critics question whether AI labs can control the systems they are building.
The story was updated on Aug. 15, 2026. On 13 August 2026, activist Wynd Kaufmyn was identified at her home in Berkeley, California; her legal case and Zhou’s FBI referral now frame the same unsettled question from opposite directions: who controls the consequences when AI development outruns judgment?
Based on
