Jacob Coxon, a former researcher at the artificial intelligence firm Anthropic, joined approximately 100 protesters outside the company’s San Francisco headquarters on Thursday, November 19, 2026. The demonstration coincided with Anthropic’s first day of trading on the Nasdaq stock exchange. Coxon resigned from his position in September to raise concerns about the pace and safety of AI development.
Coxon previously worked for several years at OpenAI before moving to Anthropic. At the time of his resignation, he stated that those developing the technology believe it poses an existential threat to humanity within this decade. Earlier in November, Coxon acknowledged he still holds equity in OpenAI, which he characterized as "blood money" given his current stance on the industry.
The protest specifically called for a ban on recursive self-improvement (RSI), a process where AI systems are used to train and improve subsequent generations of AI. Coxon told Agence France-Presse (AFP) that RSI could compress decades of research into months. He argued that if developers do not fully understand these systems, the process could lead to "highly dangerous, highly advanced AI systems by the end of 2027."
For ordinary individuals, the concrete change described by protesters involves the potential for "self-improving superintelligence" to create societal risks. Coxon and the demonstrators stated that if RSI continues, the development cycle of AI could accelerate to a point where oversight becomes difficult, potentially resulting in unpredictable systems by late 2027. This could influence future regulatory debates regarding how AI models are trained and whether certain development methods, like RSI, should be restricted by law.
The protest sets a precedent for former industry insiders publicly challenging the business models and technical methodologies of their former employers during significant financial milestones like an initial public offering. What happens next involves monitoring whether Anthropic or other AI firms respond to the calls for a ban on recursive self-improvement. No specific legislative votes or court dates related to these demands were reported. Anthropic ended its first day of trading on November 19 with a $1.8 trillion valuation, establishing its baseline as a public entity.
