The Wall Street Journal cited sources as saying that senior White House officials are intensifying their response to potential cybersecurity threats posed by the latest artificial intelligence (AI) models. Anthropic recently disclosed that its new AI model, Mythos, has an exceptionally strong ability to discover and exploit software vulnerabilities, leading the company to decide against a public release. However, a preview version of the model was provided to approximately 50 companies and organizations operating critical infrastructure, including Apple (AAPL.US), Amazon (AMZN.US), and Google, with discussions held with government officials regarding the model’s cyber capabilities.
Reports indicate that Sean Cairncross, the U.S. National Cyber Director, is spearheading the government’s response efforts. He is convening officials from various agencies to identify security vulnerabilities in critical infrastructure, strengthen government systems that could be exploited by AI, and collaborate with the Office of Science and Technology Policy and the National Security Council to ensure the government is prepared to prevent potential hacking incidents when the new model is publicly released.
It was also reported that a group of officials handling the issue held a conference call with U.S. Vice President Vance, Treasury Secretary Beson, and several top executives from the tech industry, including leaders from Anthropic, OpenAI, Microsoft (MSFT.US), Google, CrowdStrike (CRWD.US), and Palo Alto Networks (PANW.US). The discussion focused on how to address potential cyberattacks when the new model is launched.
