- November 22, 2026
- Updated 7:59 am
Anthropic’s IPO Raises Concerns Over AI Safety
- 42 Views
- admin
- November 20, 2026
- Tech Companies Technology
When Jacob Coxon, a former researcher at Anthropic, departed in September, his concerns about artificial intelligence development resonated nationwide. These concerns are situated around the rapid advancements driving public fear.
On the day Anthropic made its stock available to the public, Coxon highlighted safety issues related to AI. “It doesn’t matter whether they invest,” he remarked to Business Insider. “We have to stop RSI.” RSI stands for recursive self-improvement, a technique involving advanced AI systems training the next AI generations.
On Thursday, Coxon gathered with many outside Anthropic’s San Francisco headquarters for a protest demanding an end to RSI. This coincided with the company’s blockbuster initial public offering, which valued Anthropic at $1.7 trillion.
Protesters carried signs, listened to speeches from Coxon, former OpenAI researcher Daniel Kokotajlo, and YouTuber Rob Miles. Additionally, members from Stop AI blocked the office’s entrance later during the protest.
An independent study by the AI safety organization METR found that AI is already boosting Anthropic’s R&D efforts. Anthropic’s IPO prospectus acknowledged possible risks of allowing AI to independently design its next generation. The company is contemplating policy changes to maintain human oversight of future developments.
Coxon urged Anthropic to lobby the government and collaborate with other firms to ensure a complete ban on RSI. He stressed that only a few locations, including Anthropic’s office, could potentially implement RSI.
Prominent AI safety advocates joined the IPO protest. Both Coxon and Kokotajlo mentioned that privately, AI researchers express unease over AI’s risks. These concerns have even reached Congress, where discussions on RSI restrictions have begun.
Kokotajlo described RSI as an action reminiscent of a fictional supervillain’s plot. He emphasized the significance of addressing these issues publicly.
Oliver Habryka, who leads the company behind LessWrong, expressed relief that Anthropic will be legally required to disclose more information post-IPO. However, he remains worried about RSI, especially with AI’s increasing capabilities, particularly in biology.
David Krueger from the AI-risk nonprofit Evitable shared his belief that more public awareness of RSI would lead to widespread agreement on its risks. He commended the protesters, yet noted the challenge of reaching Anthropic employees with these protests.