The Wall Street Journal (WSJ) reported on the 11th that senior officials at the U.S. White House are moving quickly to respond to the potential threats of Anthropic's cutting-edge artificial intelligence (AI) model Mythos, which is said to have overwhelming performance in detecting security flaws.
According to the WSJ, Seán Keanecross, Director General of the U.S. Cybersecurity Directorate, has convened officials from key agencies under the Trump administration to identify security flaws in critical national infrastructure and strengthen government systems that may be vulnerable to AI-based attacks. The WSJ said, "This shows that AI risks are emerging as a top priority for the Trump administration."
The White House last week convened Vice President JD Vance, Treasury Secretary Scott Bessent, Federal Reserve (Fed) Chair Jerome Powell, and executives from major tech and finance to discuss responses to potential cyberattacks and measures to prepare online systems. Bloomberg News said the meeting was attended by Jane Fraser, Citigroup chief executive officer (CEO); Ted Pick, Morgan Stanley CEO; Brian Moynihan, Bank of America (BofA) chairman; Charlie Scharf, Wells Fargo CEO; and David Solomon, Goldman Sachs chairman.
The Trump administration's moves come as major AI corporations such as Anthropic, OpenAI, and Google race to release high-performance AI models, and the "Mythos" unveiled by Anthropic last week is seen as having pulled the trigger on the unease.
On the 7th (local time), Anthropic said it would make the "Claude Mythos Preview," which excels at detecting software security vulnerabilities, available on a limited basis only to some big tech corporations and about 40 selected institutions. Anthropic explained that "Mythos has already reached a level that surpasses most people except top-tier experts in finding and exploiting security vulnerabilities," and that access would be limited to some corporations to prevent groups or individuals with malicious intent from abusing Mythos.
OpenAI is also expected to unveil a latest AI model with similar cybersecurity capabilities in the near future.
The WSJ said Director General Keanecross is working with the Office of Science and Technology Policy (OSTP) and the National Security Council (NSC) to prepare to block dangerous hacks that may occur when the latest AI models are released. Sriram Krishnan, the White House AI policy adviser, is also reportedly involved in the response effort.
Last week, they held a video conference with big tech corporation executives including Dario Amodei, Anthropic CEO; Sam Altman, OpenAI CEO; Satya Nadella, Microsoft CEO; and Sundar Pichai, Google CEO, and leaders of cybersecurity corporations such as CrowdStrike and Palo Alto Networks also joined.
Meanwhile, Anthropic was designated a supply chain risk corporation after clashing with the U.S. Department of Defense (the Pentagon) over the scope of AI's military use. The Department of Defense is said to have used Palantir's AI-based military intelligence platform "Maven Smart System," which is equipped with Anthropic's AI model "Claude," in an airstrike on Iran.