Report says White House considers guidance to bypass Anthropic risk designation
According to Axios, US officials are drafting guidance that could allow agencies to use AI models from Anthropic despite earlier security concerns.
The White House is reportedly preparing guidance that could allow federal agencies to use AI systems from Anthropic, despite an existing risk designation.
According to Axios, the draft measure would provide a pathway to bypass supply chain concerns linked to the company. This could enable agencies to adopt newer models, including Anthropic’s recently released Mythos system.
The issue stems from earlier tensions between Anthropic and US defence authorities. The company had refused to relax safeguards related to military use, including restrictions on autonomous weapons and surveillance applications. This led to a deterioration in its relationship with the Pentagon.
The proposed guidance suggests a possible shift. It may allow agencies to proceed with procurement while broader policy disagreements remain unresolved. This reflects ongoing balancing between security concerns and access to advanced AI capabilities.
Public statements indicate some easing of tensions. US President Donald Trump recently suggested that discussions with Anthropic were progressing, following meetings between company leadership and government officials.
At the same time, divisions remain within government. Some stakeholders continue to support restrictions, while others favour re-engagement. The outcome is uncertain, and further negotiations are expected.
