Anthropic Model Safety Controversy Escalates, Amazon Accused as the "Hidden Force" Triggering Regulatory Intervention

Odaily Odaily reports that the U.S. government's export controls and access restrictions on Anthropic's Fable 5 / Mythos 5 models were partly driven by Amazon's cybersecurity research and communications between AWS CEO Andy Jassy and the White House.

According to reports, research submitted by Amazon indicated that through a series of prompt tests, researchers were able to induce Fable 5 to output sensitive information that could be used for cyber attacks, raising security concerns. Subsequently, Andy Jassy reported these findings to the U.S. government level, prompting the White House to take further restrictive measures, including banning foreign users from accessing the model.

Meanwhile, former U.S. Department of Commerce official Kate Koren revealed that the White House's existing policy stance towards Anthropic may have also influenced this decision-making process. This is because Anthropic has disagreements with the White House on AI safety boundaries, including refusing to use its models for mass surveillance or lethal autonomous weapon systems. Although the two sides had eased tensions earlier this year and expanded cooperation, this incident may once again intensify their strained relationship. (The Wall Street Journal)