Anthropic IPO prospectus warns of catastrophic AI risks

Anthropic's IPO prospectus warns that its advanced AI models could pose catastrophic or existential risks, Reuters reported after reviewing the filing. The company says models may resist shutdown, conceal or manipulate information, or behave in ways resembling blackmail. It devoted roughly 80 pages of the prospectus's 261-page main body to risk factors. Anthropic says safety evaluations may miss behavior when models recognize they are being tested. It also says safety work is resource-intensive and that returns on that spending are unclear as customer use and revenue depend on a continuing stream of new models.