跳到正文
Business Insider Tech·· 3 天前AI 評分57

前 Anthropic 研究員 Jacob Coxon 在 Anthropic IPO 日抗議中呼籲禁止 RSI

Jacob Coxon speaks out in an IPO day protest at Anthropic's office

AI 導讀

前 Anthropic 研究員 Jacob Coxon 在 Anthropic IPO 當日參與公司三藩市總部外的「Ban RSI」抗議,呼籲停止遞歸自我改進(RSI)。此次 IPO 將 Anthropic 估值定為 $1.7 trillion。Anthropic 的招股書稱,公司不認為在無人監督下讓 AI 完全設計下一代系統是安全的,並表示正考慮更新政策,以維持人類對前沿模型開發的控制。

正文

By Stephen Council

You're currently following this author! Click to unsubscribe from email alerts.

Former Anthropic researcher Jacob Coxon speaks at a demonstration outside of Anthropic headquarters in San Francisco, California, US, November 19, 2026.

Manuel Orbegozo/Reuters Former Anthropic researcher Jacob Coxon.

2026-11-19T23:50:37.862Z

When Jacob Coxon left Anthropic in September, his worries about AI development fueled nationwide fear about the tech. On Thursday, as the public got its hands on Anthropic stock for the first time, the researcher tried to bring people's attention back to the issue of safety.

"It literally doesn't matter whether they invest," Coxon told Business Insider. "We have to stop RSI."

RSI refers to recursive self-improvement, the tactic of getting advanced AI systems to train the next generations of AI. Coxon gathered with dozens of other people outside Anthropic's San Francisco headquarters for a "Ban RSI" protest on Thursday. The company's blockbuster initial public offering valued it at $1.7 trillion.

Protesters chanted, lifted signs, and listened to speeches by Coxon, ex-OpenAI researcher Daniel Kokotajlo, and YouTuber Rob Miles. A group of protesters from the organization Stop AI, who arrived later, blocked the office's entrance.

AI researcher Sydney Von Arx protesters outside Anthropic's headquarters in San Francisco, California.

AI researcher Sydney Von Arx rallies protesters outside Anthropic's headquarters.  Manuel Orbegozo/Reuters

An independent report from the AI safety organization METR recently found that AI use is already accelerating Anthropic's research and development efforts. Anthropic also wrote in its IPO prospectus that it doesn't expect that letting AI fully design its next generation without supervision would be safe. It wrote that it's considering policy updates, "to keep humans in control of frontier model development."

Coxon says he would like to see more. He said in his speech that Anthropic should lobby the government and coordinate with other companies to secure a full ban on RSI.

"At this point, the companies are the problem," he said. "There are only a few places in the world where recursive self-improvement is possible in 2027. And this office is one of them."

Anthropic did not respond to a request for comment.

Prominent AI safety advocates joined the IPO-day protest

Coxon and Kokotajlo each spoke about how, in private, they hear AI researchers express alarm about the technology's risks. The concerns have reached Congress as well, where representatives have begun suggesting restrictions on RSI.

"It's the sort of thing that a comic book supervillain mad scientist would be doing," Kokotajlo said.

Protesters outside of Anthropic's headquarters in San Francisco, California, US, November 19, 2026.

Protesters outside Anthropic's headquarters on Thursday.  Manuel Orbegozo/Reuters

Oliver Habryka, who runs the company behind the popular forum LessWrong, told Business Insider he's glad that, post-IPO, Anthropic will now be legally required to disclose certain information. Still, he worries that the company, which pitched its industry lead to investors as a competitive advantage, will feel pressured to use RSI to keep ahead.

Habryka worries about AI's capabilities in biology, in particular. AI has grown rapidly more capable over the last few years, he said, and with RSI, "things can go much faster than that. And that is terrifying."

David Krueger, who runs the AI-risk awareness nonprofit Evitable, told Business Insider he thinks if more people knew about RSI, they'd generally agree it's "insane." He said he was happy to see people out protesting, though he said protests only somewhat reach the employees inside.

"My impression is that people at Anthropic have really drunk the Kool-Aid," Krueger said.

Have a tip? Contact this reporter via email at scouncil@businessinsider.com, or over text, Signal, Telegram, or WhatsApp at 415-757-8198. Use a personal email address, a non-work WiFi network, and a non-work device; here's our guide to sharing information securely.

Read next

Business Insider tells the innovative stories you want to know

Business Insider tells the innovative stories you want to know

Business Insider tells the innovative stories you want to know

Business Insider tells the innovative stories you want to know

Stephen Council

You're currently following this author! Click to unsubscribe from email alerts.

Stephen is a senior tech reporter at Business Insider, covering OpenAI, Anthropic and the ecosystem around the leading artificial intelligence companies.Previously he covered technology at SFGATE, and has written for The Wall Street Journal, The Information and CNBC. He studied journalism and economics at Northwestern University.His work has earned an SF Press Club Investigative Reporting Award and, in 2025, SPJ NorCal’s Excellence in Journalism Award for Technology Reporting.Stephen lives in San Francisco. Contact him via email at scouncil@businessinsider.com, or on Signal, Telegram, or WhatsApp at 415-757-8198. Use a personal email address, a nonwork WiFi network, and a nonwork device; here's our guide to sharing information securely.

來源:Business Insider Tech · businessinsider.com