Anthropic IPO filing warns AI harm could exceed insurance capacity

Anthropic’s public filing for a planned initial public offering warns that a severe incident involving advanced artificial intelligence could cause harm on a scale that exceeds the capacity of individual insurers or the insurance industry. The company says such outcomes are possible risks, not events that have occurred, and could bring legal, regulatory and reputational consequences. The Form S-1, filed October 26, describes catastrophic or existential risks if advanced systems are developed or used without adequate safeguards, oversight and security. Anthropic lists possible misuse by malicious actors, including help with dangerous weapons, as well as systems operating with a high degree of autonomy. The filing says safeguards could be bypassed or fail to prevent harm. Anthropic also warns that a damaging incident could affect customers and the public, expose the company to liability and regulatory costs, erode trust and reduce demand. The filing says losses could exceed what an insurer, or insurers collectively, are able or equipped to cover. These are forward-looking disclosures in the registration statement, rather than evidence of an actual catastrophic incident. The...