Full Breakdown
NSA Utilizes Anthropic's Mythos AI Amid Ongoing Dispute with Pentagon
4/20/2026, 6:08:44 AM
NSA's Use of Mythos AI Tool
The United States National Security Agency (NSA) has begun utilizing Anthropic's Mythos Preview AI tool, despite a formal supply-chain risk designation issued by the Pentagon against the company. Reports from Axios indicate that the NSA is among approximately 40 organizations granted access to Mythos, which was introduced in early April 2026. Sources suggest that the tool is being used more extensively within the agency, primarily for identifying cybersecurity vulnerabilities.
Background of the Dispute
The relationship between Anthropic and the U.S. government has been fraught since February 2026, when President Donald Trump ordered all federal agencies to cease using Anthropic's services. This directive followed Anthropic's refusal to comply with Pentagon demands for unrestricted access to its AI models for military applications, including potential use in autonomous weapons systems. In response, Anthropic filed lawsuits against the Department of Defense, challenging its classification as a "supply chain risk," which effectively blacklisted the company from military contracts.
Key Developments
Despite the ongoing legal battles, discussions between Anthropic's CEO Dario Amodei and White House officials, including Chief of Staff Susie Wiles, have taken place. These meetings aim to explore a resolution that would allow the government to utilize Mythos while addressing safety concerns. The White House described the recent meeting as "productive and constructive," although President Trump claimed to be unaware of it when questioned by reporters.
Implications of Mythos AI
Anthropic's Mythos is characterized as a powerful general-purpose language model with advanced capabilities in cybersecurity tasks. Experts have noted that its ability to identify and exploit vulnerabilities could pose significant risks if misused. The UK’s AI Security Institute has evaluated a preview version of Mythos, finding it substantially more capable of offensive cyber operations than previous models. This dual-use nature of the technology raises concerns about its potential for both defensive and offensive applications.
Criticism and Opposition
Critics within the Pentagon argue that Anthropic's refusal to grant unrestricted access to its models demonstrates a lack of reliability for military needs. Some officials express concern that the company's stance could hinder national security efforts. Conversely, others advocate for a resolution that would enable the government to leverage Anthropic's advanced tools to enhance cybersecurity defenses.
Conflicting Reports and Gaps
While some sources indicate that the NSA is using Mythos extensively, the specifics of its application remain unclear. Additionally, the legal status of Anthropic's classification as a supply-chain risk continues to evolve, with one court granting a preliminary injunction against the designation while another denied a similar request.
What's Next
The ongoing negotiations between Anthropic and the U.S. government will likely focus on how various agencies can safely engage with Mythos. As the urgency surrounding AI technology grows, the outcome of these discussions may significantly impact both national security and the future of AI development in the United States.
Verbatim Quotes
- “It would be grossly irresponsible for the U.S. government to deprive itself of the technological leaps that the new model presents,” — Source, Axios
- “The paradox is that the same government that blacklisted Anthropic now wants access to its most powerful model.” — Source, Axios
- “reveals a lot more vulnerabilities” — Jamie Dimon, CEO of JPMorgan Chase
- “Second, that even in situations of political tension between companies and governments, technological dependence often overshadows political considerations.” — Source, Axios
