NSA Readies Anthropic’s Mythos for Cyber Operations Applications

A report from The Financial Times states that Anthropic has deployed around six engineers to the National Security Agency (NSA) to help in leveraging its cybersecurity AI model, Mythos, based on information from anonymous sources.

These engineers are thought to be aiding the intelligence agency in applying Mythos for particular operations, although it remains unclear if they or the model are involved in the agency’s hacking efforts.

The NSA is tasked with gathering intelligence via wiretaps, undersea cables, corporate collaborations, and other covert techniques, in addition to carrying out offensive cyber actions against foreign threats.

Moreover, the FT’s report expands upon earlier insights from Axios, which revealed in April that the NSA was using Mythos despite a federal restriction against deploying Anthropic’s technology. This restriction originated from the Department of Defense’s designation of Anthropic as a “supply chain risk” after the company opted not to permit the government to use its models for extensive domestic surveillance and autonomous weaponry.

This latest information emerges as governments rapidly seek access to Mythos, which Anthropic asserts it had to limit due to concerns that its cybersecurity capabilities could be misappropriated for discovering vulnerabilities and executing hacks.

When approached by TechCrunch, an NSA representative declined to confirm or deny the report, and Anthropic did not respond to requests for comment.