Anthropic Proposes Deal to Commerce Secretary to Lift Ban on Advanced AI Models
The AI company is pledging closer White House collaboration and faster security issue resolution to end export controls on its 'Mythos' and 'Fable' models.
Anthropic is engaging in discussions with Commerce Secretary Howard Lutnick, proposing increased cooperation with the White House to address security concerns that led to a U.S. ban on its powerful "Mythos" and "Fable" artificial intelligence models. Sources familiar with the situation stated that the artificial intelligence company has committed to enhancing its communication with the Trump administration and expediting the resolution of any future security vulnerabilities.
Negotiations are reportedly progressing positively, although a definitive timeline for the removal of the export controls remains uncertain. Anthropic's chief compute officer, Tom Brown, and head of public policy, Sarah Heck, are leading the talks with administration officials. Representatives for Anthropic and the Commerce Department have not yet commented on the ongoing discussions.
Crackdown Follows Safety Concerns
Anthropic initially dispatched top executives to Washington D.C. late last week after the Trump administration imposed export controls on "Mythos" and "Fable." This action followed warnings from Amazon CEO Andy Jassy, who informed the administration that researchers had discovered ways to bypass the AI models' safety guardrails. In response, Anthropic opted to temporarily withdraw the models, stating it was necessary to comply with the export regulations.
The issue gained attention at the G7 Summit in France, where President Trump commented that talks with Anthropic were proceeding "fine." Meanwhile, Anthropic CEO Dario Amodei urged world leaders to maintain a unified approach to AI regulation.
An Anthropic spokesperson previously stated that both parties are working swiftly toward a resolution, emphasizing the company's commitment to collaborating with the administration to protect U.S. critical infrastructure and maintain American leadership in cyber defense. Ironically, the White House's action against Anthropic occurred shortly after Amodei publicly advocated for U.S. government authority to shut down potentially dangerous frontier AI models.
White House officials reportedly expressed frustration that Anthropic and Amodei initially characterized the safety flaw as a "narrow" issue, despite previous warnings from the company about potential AI risks. A senior administration official indicated that the White House had received similar warnings from nearly a half-dozen other companies prior to taking action. The official suggested that the situation could have been avoided if Anthropic had promptly addressed or paused access to the models rather than dismissing the concern.