NSA Adopts Anthropic's Mythos AI Model Amid Easing Tensions

The National Security Agency (NSA) has commenced use of Anthropic's Mythos Preview, an advanced artificial intelligence model, according to sources familiar with the matter. This development follows a period of contention between Anthropic and the Pentagon regarding the deployment of AI technologies for military purposes.

Anthropic introduced Mythos Preview in early April, describing it as a general-purpose language model with notable capabilities in computer security tasks. The model has been selectively released to approximately 40 organizations, including the NSA, as part of a controlled initiative aimed at enhancing cybersecurity measures.

The adoption of Mythos Preview by the NSA comes shortly after Anthropic CEO Dario Amodei met with White House Chief of Staff Susie Wiles and other officials to discuss the AI model's potential applications. A White House spokesperson characterized the meeting as "productive and constructive," emphasizing the administration's interest in collaborating on AI advancements while ensuring safety and security.

This engagement marks a significant shift from earlier tensions, where the Pentagon had labeled Anthropic a "supply chain risk" following the company's refusal to modify certain safeguards for military use. The recent developments suggest a move towards reconciliation and cooperation between Anthropic and U.S. government agencies in the realm of artificial intelligence and cybersecurity.