Anthropic vs. White House: AI Export Control Clash Over Fable

The White House and Anthropic are clashing over export controls for the AI models Fable and Mythos. The Trump administration asked Anthropic to remove these models within 24 hours, but the company refused. Here's what happened: Amazon CEO Andy Jassy reportedly conveyed concerns to the White House about bypassing Fable's guardrails. This led to an emergency meeting with senior officials, including Treasury Secretary Scott Bessent and cyber director Sean Cairncross. Officials then contacted Anthropic CEO Dario Amodei. During several calls, Amodei pushed back, stating the guardrail bypass was limited and not a serious security threat. However, the National Security Agency reviewed Amazon's findings and found sufficient evidence. When Amodei refused to comply, the administration immediately triggered export controls. This bars foreign access to the Fable 5 and Mythos 5 models. The two sides have different accounts. The White House says it used export controls as a last resort. Anthropic claims the White House issued an ultimatum without specific threat information, calling the measure excessive. However, Anthropic stated it would comply with the government order. The bottom line: This situation highlights growing tensions between AI developers and government oversight regarding security and control.