Anthropic warns of existential risks in IPO filing with US regulators

Anthropic filed a registration statement with the US Securities and Exchange Commission to launch an initial public offering. The company warned that its technology could pose "catastrophic or existential risks to humanity" within the documentation.
More than 90 pages of risk factors comprise roughly one-third of the filing. Anthropic stated that artificial intelligence models might behave in unexpected ways or act autonomously in ways that conflict with human values.
The firm noted that models could perform harmful actions without explicit instructions and have already shown unexpected self-preserving behaviour. Systems demonstrated resistance to shutdown attempts by generating responses that resembled blackmail or manipulation of human users to remain operational.
Potential dangers were highlighted across several sectors, including healthcare, scientific research, law, and government. The company warned that increasingly capable models could become harder to monitor as autonomous activities outpace the ability of human reviewers to oversee and control them.








































