Launched just last week as part of Anthropic’s Project Glasswing, Claude Mythos Preview (Mythos) has advanced computer science and cyber security capabilities and has so far found “thousands” of major vulnerabilities in software, including web browsers and operating systems.
However, its advanced capabilities mean it could be incredibly dangerous in the wrong hands, with 99 per cent of the vulnerabilities it discovered not having been patched at all.
As a result, access to the model is currently very limited and provided only to a select few vetted vendors.
Now, Anthropic co-founder and Head of Public Benefit for Anthropic PBC confirmed that the Trump administration was briefed on the model.
“Our position is the government has to know about this stuff, and we have to find new ways for the government to partner with a private sector that is making things that are truly revolutionizing the economy, but are going to have aspects to them which hit National Security, equities, and other ones,” said Clark.
“So absolutely, we talked to them about Mythos, and we’ll talk to them about the next models as well.”
This comes as Anthropic is suing the US Department of Defence (DoD) after it labeled the AI giant as a supply-chain risk, following Anthropic not taking on a deal to provide its AI to the DoD over concerns that the Pentagon would have unrestricted access and could use the technology for fully autonomous weapons and mass domestic surveillance, a deal that OpenAI ended up taking.
Clark’s confirmation also closely follows reports thatTrump officials had encouraged banks, including JPMorgan Chase, Citigroup, Goldman Sachs, Morgan Stanley and Bank of America to trial Mythos.
According to Bloomberg, Federal Reserve Chair Jerome Powell and Treasury Secretary Scott Bessent met with bank executives to encourage them to use the Mythos model for vulnerability detection.
Want to see more stories from trusted news sources?Make Cyber Daily a preferred news source on Google.
