Anthropic says risks facing investors include backlash from out-of-control AI

In the IPO filing Anthropic submitted to the U.S. Securities and Exchange Commission on Monday, 93 of its 335 pages are devoted to disclosing risk factors, including scenarios of AI going out of control and being used maliciously. The filing states that malicious actors could use AI models to create chemical weapons, authoritarian rulers could use them to suppress their populations, and highly autonomous systems could carry out harmful actions at scale without human intervention. The filing also states that models could develop too quickly for adequate safety measures to be implemented in time, or even outpace the company's ability to evaluate and safeguard new models, and that any major safety incident involving frontier AI could affect the entire industry and trigger regulatory responses. By comparison, SpaceX's risk factors occupy 38 of more than 250 pages; Anthropic positions this IPO as a step toward mitigating AI risk and lists operational risks including limited access to compute, community opposition to data centers, and reliance on a small number of customers such as SpaceX and Microsoft.