Anthropic Engineers Deploy AI Cyber Tool 'Mythos' at NSA Amid Controversy and Global Expansion
June 5, 2026
Anthropic has embedded approximately six forward-deployed engineers inside the NSA to deploy Mythos for offensive cyber operations, according to the Financial Times.
The engineers are said to help the NSA leverage Mythos for certain applications, though it remains unclear whether the model is actively used in hacking operations.
Experts warn Mythos could be used to infiltrate networks of adversaries such as China or Iran, highlighting concerns about AI-driven offensive capabilities and dual-use risks.
Anthropic recently expanded Mythos access to additional companies under security-focused arrangements, despite earlier disputes with the Pentagon.
The expansion and AI cybersecurity developments come as governments deepen involvement and as AI firms compete, including OpenAI's related cybersecurity-focused efforts.
Anthropic has filed for an IPO that could value the company at over $1 trillion, with rivals like OpenAI releasing similar capabilities since Mythos’s debut.
Governments, financial institutions, and IT firms remain concerned about Mythos’s ability to detect and exploit software vulnerabilities.
Mythos, unveiled in early spring, is described as highly dangerous and capable of identifying security vulnerabilities and exploitable attack paths.
The NSA deployment comes amid a legal dispute with the Pentagon over restrictions on government use of Claude AI models for warfighting and mass surveillance, which the Pentagon has flagged as a supply-chain risk.
Anthropic’s dual stance—fighting the Pentagon in court while deploying at the NSA—raises questions about private AI firms’ role in state-sponsored cyber operations and potential proliferation of Mythos.
Anthropic initially rolled out Mythos to select U.S.-based organizations in April and is expanding to about 150 organizations across 15 countries, with plans to add more countries and beneficiaries.
The expansion centers on broader access to Mythos amid concerns from governments and industry about its ability to identify and exploit security vulnerabilities.
Summary based on 8 sources
Get a daily email with more Tech stories
Sources

TechCrunch • Jun 5, 2026
NSA said to be readying Anthropic’s Mythos for use in cyber operations
Investing.com • Jun 4, 2026
Anthropic embeds engineers at NSA to deploy Mythos AI model - FT
MLQ.ai • Jun 5, 2026
MLQ.ai | AI for investors
Firstpost • Jun 5, 2026
Anthropic embeds engineers at US security agency to deploy Mythos AI: Report