{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreifl25xq6aljgxrsogwozxvgas3d32csrbhfy54iz4t5kn7gqazidu",
    "uri": "at://did:plc:mn7j4jclhui5ehnaibt6jfba/app.bsky.feed.post/3mnk2yme64sb2"
  },
  "coverImage": {
    "$type": "blob",
    "ref": {
      "$link": "bafkreihnpp2jacnov45c47ydaiutqijpppo5lgkxacs33zmucxnmugps4i"
    },
    "mimeType": "image/jpeg",
    "size": 257654
  },
  "description": "Anthropic has placed about half a dozen “forward-deployed engineers” inside the US National Security Agency (NSA) to help use Mythos.",
  "path": "/nsa-reportedly-deploying-anthropics-mythos-ai-months-after-pentagon-dispute/",
  "publishedAt": "2026-06-05T11:48:24.000Z",
  "site": "https://www.techloy.com",
  "tags": [
    "Anthropic",
    "Claude",
    "officially label",
    "began",
    "blocked",
    "reports",
    "Mythos",
    "Claude Mythos Found Thousands of Holes in U.S. Bank Systems. Now Wall Street Is ScramblingAnthropic’s restricted AI model has Wall Street banks patching at speeds they’ve never attempted before, with a 6-month window before Chinese AI catches up.TechloyDamilare Odedina",
    "emergency discussions",
    "Pentagon Moves to Replace Anthropic AI After Dispute Over Military UseThe Trump administration has labelled Anthropic a “supply-chain risk,” putting its existing and future defence contracts in jeopardy.TechloyOgbonda Chivumnovu"
  ],
  "textContent": "It wasn’t long ago that Anthropic, the AI company behind Claude, sued the US government, challenging the Pentagon’s decision to officially label the company a “supply chain risk.” The messy legal dispute began when a deal between Anthropic and the Department of War (DOW) over the use of Anthropic’s AI systems fell apart.\n\nThe company reportedly refused to allow the DOW to use its AI systems because it demanded contractual “red lines” that would block the military from using its technology for mass domestic surveillance or fully autonomous lethal weapons. Although, in March, a judge blocked the DOW’s move to label Anthropic a “supply chain risk,” citing free speech violations.\n\nBut now, the _Financial Times_ reports that Anthropic has placed about half a dozen “forward-deployed engineers” inside the US National Security Agency (NSA) to help deploy one of its most powerful models, Mythos, for offensive cyber operations.\n\nIt’s pretty ironic because the US government once seemed adamant about blacklisting Anthropic AI products for government agencies yet is now reportedly engaging with the company.\n\nBut this could make sense when you look at all the controversy and discussion surrounding Mythos since its preview earlier this year. The AI model has been dubbed by many as the most powerful AI model Anthropic has ever released, mostly because of its advanced ability to identify and exploit zero-day vulnerabilities autonomously at an unprecedented level.\n\nClaude Mythos Found Thousands of Holes in U.S. Bank Systems. Now Wall Street Is ScramblingAnthropic’s restricted AI model has Wall Street banks patching at speeds they’ve never attempted before, with a 6-month window before Chinese AI catches up.TechloyDamilare Odedina\n\nThis caused panic across the tech world, leading Anthropic to limit access to the model to only a handful of companies and government organisations. It even prompted US officials to hold emergency discussions about the cyber preparedness of major banking institutions and large organisations in the event a model like this fell into the wrong hands.\n\nThat concern could also explain why the NSA may want access to the technology itself. The logic appears to be that understanding how an AI model can launch sophisticated cyberattacks could also help governments prepare stronger defences against them. Even one person close to Anthropic, as reported by the _Financial Times_ , shared similar sentiments, saying, “The best way to build a good defence is to build a good attack…. If [Mythos] is not used to build attack agents, adversaries will find a way to do it.”\n\nNeither Anthropic nor the NSA has publicly commented on the reported partnership. But it does not look like the NSA will be the only organisation getting access to Mythos at this time. This week, reports suggest Anthropic plans to announce the distribution of Mythos to 150 organisations across 15 countries, vastly expanding access beyond the US and UK.\n\nMany experts and researchers have voiced concerns about Mythos’ capabilities and the danger it could pose if it falls into the wrong hands. It looks like the US government may already be taking a proactive approach to address these concerns, as President Trump on Tuesday signed an executive order creating a voluntary framework for AI companies to submit new models for security reviews before public release. The order also directs federal agencies to evaluate the cyber capabilities of AI systems.\n\nAnthropic built a powerful AI model for cybersecurity. Now, it is reportedly helping the NSA use it offensively while simultaneously fighting the government in court. To some, that might seem hypocritical. But considering Mythos is viewed as one of the most powerful AI models currently, and everyone seemingly wants access to it, it makes sense why the NSA would want to be among them.\n\nPentagon Moves to Replace Anthropic AI After Dispute Over Military UseThe Trump administration has labelled Anthropic a “supply-chain risk,” putting its existing and future defence contracts in jeopardy.TechloyOgbonda Chivumnovu",
  "title": "NSA Reportedly Deploying Anthropic's Mythos AI Months After Pentagon Dispute",
  "updatedAt": "2026-06-05T11:48:25.718Z"
}