A San Francisco courtroom is about to hear arguments that blocking a company's doors might be justified if the alternative is human extinction. Wynd Kaufmyn, a retired City College of San Francisco professor and activist with the group Stop AI, stands trial for chaining shut the entrance to OpenAI's headquarters in February 2025. She faces up to a year in prison on four counts of misdemeanor trespassing, but her defense strategy hinges on something far larger than the charges themselves: whether her actions were necessary to prevent imminent catastrophic harm from artificial general intelligence (AGI), a hypothetical AI system that would match or exceed human intelligence across all domains.

What Is the "Necessity Defense" in an AI Context?

Kaufmyn's legal team has filed a motion to use what's known as a necessity defense, which allows defendants to argue that breaking the law was justified to prevent greater harm. The judge has permitted her to present this argument, meaning a jury will hear testimony about whether stopping a machine can excuse breaking the law. This is unprecedented territory. The defense will argue that Kaufmyn's blockade was a reasonable response to what she and other activists view as reckless development of superintelligent systems.

On the witness stand, UC Berkeley AI professor Stuart Russell is expected to testify about whether Kaufmyn's actions were justified. The defense has also subpoenaed Sam Altman, OpenAI's chief executive officer, as a non-party witness, though his lawyers have filed a motion to quash the subpoena. The trial will essentially put the existential risk argument itself on trial, forcing a jury of ordinary people to weigh claims about AI safety against the legal question of trespassing.

Why Are Activists Comparing This to the Civil Rights Movement?

Kaufmyn's supporters are drawing explicit parallels to Rosa Parks, the civil rights icon arrested in 1955 for refusing to move to segregated seating on a Montgomery bus. David Krueger, an AI risk expert and CEO of Evitable, suggested that Kaufmyn could become "the Rosa Parks of AI risk" if the case catalyzes a broader civil disobedience movement.

"Governments have been asleep at the wheel and activists like Wynd have stepped up to face this crisis. This case marks the first time a court will get to hear some of the copious evidence that AI companies are gambling with our lives. It could also mark the beginning of a much larger campaign of civil disobedience, like what we saw with the civil rights movement. In ten years time, we may look back on Wynd as the Rosa Parks of AI risk," said David Krueger.

David Krueger, AI Risk Expert and CEO of Evitable

The comparison reflects a conviction among some AI safety advocates that the stakes are existential. Stop AI and allied groups have been actively protesting against OpenAI and other AI companies, demanding an immediate halt to frontier AI research that could lead to AGI and superintelligence. Activists argue both states are dangerous and capable of leading to "human extinction" as AI systems would far exceed human capabilities.

What Evidence Will the Court Hear About AI Risk?

The trial will center on whether there is genuine, imminent harm from AI development that justifies civil disobedience. Krueger addressed the core claim: that AI companies cannot reliably determine whether their systems are safe. He emphasized a concept called "recursive self-improvement," where AI systems improve themselves without human oversight, which experts widely acknowledge as a probable point of no return.

Krueger

"AI companies are not able to reliably determine whether their systems are safe with any reasonable level of confidence. Despite this, they've made clear they intend to keep racing to build and deploy more and more powerful AI, even putting AI itself in charge of that. This process, known as recursive self-improvement, is widely acknowledged as a probable point of no return," explained David Krueger.

David Krueger, AI Risk Expert and CEO of Evitable

When pressed on the scale of the alleged danger, Krueger made a stark comparison. He noted that extinction may seem abstract, but the potential harm dwarfs previous genocides. He argued that pointing out such risks is not harsh if one takes the existential threat seriously.

How Are AI Safety Advocates Framing the Moral Case?

Stop AI and allied activists maintain that they believe in nonviolent civil disobedience as a form of protest against AI companies that have received "very little or gentle criticism for their actions." The group's position reflects a philosophy rooted in moral urgency rather than legal technicality. Kaufmyn herself stated that she would testify about the catastrophic danger AI poses to humanity, arguing that the technology is being developed by people who know the existential risk and are racing forward anyway.

"On Friday, I will testify against the catastrophic danger AI poses to humanity. This technology is being developed by people who know the existential risk and are racing forward anyway. They care more for their own fortunes than for the good of humanity," said Wynd Kaufmyn.

Wynd Kaufmyn, Activist and Retired Professor at City College of San Francisco

When asked whether activists believe in leaving room for dialogue and conversion, Krueger acknowledged the tension. He suggested that an easy path for AI company employees would be to quit their jobs and join the protest movement, but he framed the urgency as leaving little room for gradualism.

Steps to Understanding the Trial's Broader Implications

• Legal Precedent: This is the first time a U.S. court will hear a necessity defense based on existential AI risk, potentially setting a precedent for how courts evaluate claims about AI safety and imminent harm.
• Expert Testimony: The trial will feature testimony from prominent AI safety researchers, including UC Berkeley's Stuart Russell, lending academic credibility to existential risk arguments in a legal setting.
• Jury Decision: A jury of ordinary citizens, not AI experts, will ultimately decide whether blocking a company's doors was justified to prevent AI catastrophe, making this a test of how the general public weighs existential risk claims.
• Activist Movement: If Kaufmyn is acquitted or receives a light sentence, it could embolden a broader civil disobedience campaign similar to the civil rights movement, as supporters have already drawn historical parallels.
• Corporate Accountability: The subpoena of Sam Altman signals that activists are willing to put AI company leadership directly on the stand to defend their safety practices and development timelines.

The trial represents a collision between two worldviews: one that sees AI development as a manageable engineering challenge requiring oversight, and another that views it as an existential threat requiring immediate cessation. For the first time, a jury will be asked to weigh these competing narratives not in a policy forum or academic conference, but in a courtroom where the verdict could carry real consequences for both Kaufmyn and the broader AI safety movement.