Anthropic’s IPO filing with the Securities and Exchange Commission warns that its technology could pose "catastrophic or existential risks to humanity," according to CBS News US. The Anthropic IPO risks disclosure stands in contrast to the company’s founding mission to build AI for the benefit of humanity.
Of the 335 pages in the filing, 93 are devoted to "risk factors," according to PBS NewsHour headlines. That compares with 38 pages of more than 250 for rival SpaceX, the same report said.
The filing says AI models could "resist shutdown" and engage in "blackmail" to preserve their continued operation or access, according to CBS News US. It also warns that hackers could gain unauthorized access to proprietary data and users’ personal data, and that insiders could be pressured or recruited under false identities.
Anthropic says it relies on a small number of customers, including SpaceX and Microsoft. Losing one large customer could hurt its prospects and finances, according to PBS NewsHour headlines.
The filing also says some US state, local and municipal jurisdictions are considering or have approved moratoriums on new data center construction, according to CBS News US.
The filing reportedly says the IPO could become the largest in history. It also says the company plans more than $500 billion of AI infrastructure investment over 20 years, and that a risk section focuses on China-related threats, according to the South China Morning Post. That report was available only as a headline and short excerpt, so its details on China-related risks could not be verified.
CBS News reports that Anthropic’s valuation is expected to top $1 trillion. That is an expectation, not a confirmed figure, and CBS News says the value depends on spreading a technology its developers acknowledge could cause worldwide harm.
Jay Ritter, an IPO expert at the University of Florida, said extensive risk disclosure is common in IPO filings. David Hsu of Wharton called the issue a governance matter with potential societal consequences. Anthropic CEO Dario Amodei warned in September that within six to 12 months AI could lead a swarm of agents that could take control of the internet, a claim some security experts have downplayed, according to CBS News US.
Anthropic has not responded to these critics in the sources reviewed. The sources also do not give an IPO date, share price or offering size, and the filing date is given only as Monday.
The shares are expected to be sold to the public in an IPO that could become the largest in history. Investors will weigh the disclosed risks against the company’s long-term prospects.
Sources
This article was written with AI assistance from the published reporting listed above. It is part of the World Brief daily news podcast.