Anthropic proposes courts block dangerous AI models — Channel NewsAsia

In the United States, Anthropic proposed that California Governor Gavin Newsom allow government officials to go to court to seek to block the release of artificial intelligence models if testing finds that they pose a serious risk to society. Channel NewsAsia reports, citing the company's letter dated October 15.

Proposals for California

An Anthropic representative confirmed the authenticity of the letter. The California governor's office did not provide an immediate comment on the initiative.

The company submitted its recommendations after Newsom formed a group of artificial intelligence experts in September. Its task is to prepare proposals on AI safety requirements.

More current news is available on the UA.News Telegram channel Telegram.

Instead of a complete shutdown

Anthropic also opposed a requirement to build so-called kill switches into models that could completely stop a system's operation if the company loses control over its actions. In the company's view, targeted measures would be more effective: shutting down a problematic AI agent or terminating a specific client's access to the model.

In September, a bipartisan group of U.S. senators discussed a bill that could give the government the authority to block the release of individual AI models deemed dangerous. Under this approach, companies could challenge such a decision in federal court. After the election recess, Congress did not advance the proposal further.

In July, members of the U.S. House of Representatives introduced the AI Kill Switch Act, which provides officials with the right to require AI companies to stop models that pose a risk to human life or the economy. Earlier, Donald Trump's administration ordered Anthropic to restrict foreign access to its cyber-capable Mythos 5 and Fable 5 models over national security concerns, but the U.S. Department of Commerce later rescinded the order.