Trump Administration Orders Export Controls on Anthropic AI Models
The Trump administration imposed export controls on Anthropic's latest artificial intelligence models, Mythos and Fable, on a Friday. This action required Anthropic CEO Dario Amodei to block foreign access to the models, which Anthropic then pulled entirely. The decision followed a 24-hour period where officials pressed the company over security concerns, particularly regarding the models' guardrails, highlighting the White House's challenges in regulating advanced AI.
The Trump administration issued export controls against AI developer Anthropic on a Friday, directing CEO Dario Amodei to block foreign access to the company's newest models, Mythos and Fable. In response, Anthropic fully withdrew access to both AI models.
This move by the administration stemmed from a frantic 24-hour effort to persuade Anthropic to voluntarily pull the newly released artificial intelligence models. Administration officials held concerns that the models posed security risks, with particular doubts arising about the robustness of their guardrails, despite Anthropic's earlier assurances of safety.
Concerns escalated on Thursday, two days after one model, Fable, was publicly released, when Amazon CEO Andy Jassy conveyed to the White House his worries about the ability to bypass the model's guardrails. Amazon, an investor in Anthropic, provided this feedback after an administration request.
By Friday morning, the issue reached the highest levels of the White House. Treasury Secretary Scott Bessent, White House Cyber Director Sean Cairncross, Chief of Staff Susie Wiles, and other senior officials convened to discuss the model and the administration's response.
Multiple tense calls ensued between Anthropic CEO Dario Amodei and various senior administration officials, including Bessent, Cairncross, and Commerce Secretary Howard Lutnick. Other officials involved in some of these calls included Undersecretary of Commerce for Industry and Security Jeffrey Kessler, White House staff secretary Will Scharf, White House deputy chief of staff Richard Walters, and assistant to the president for policy Walker Barrett. Amodei reportedly defended the models' guardrails, arguing that any identified bypasses were specific and did not present the same risk as a broader 'jailbreak' that would negate all safety measures.
This event underscores the challenges faced by the White House in regulating rapidly developing artificial intelligence technologies that are seen as potentially dangerous.
(Source: Business Insider)
Advertisement
AdSense slot • inline
