Anthropic Briefs US Government on Mythos Model

Anthropic co-founder Jack Clark confirmed that the company briefed the Trump administration on its new Mythos model. The model was assessed to be dangerous enough, due to its powerful cybersecurity capabilities, that it was not released to the public.

Clark explained why the company continues to engage with the government, saying that on matters related to national security, private companies and the government need to cooperate in new ways. He stated that Anthropic will continue talking with the government not only about Mythos but also about subsequent models.

Meanwhile, Anthropic filed a lawsuit in March in response to the DOD classifying the company as a supply chain risk. The company has clashed with the Pentagon over issues surrounding the military's demand for unrestricted access to Anthropic AI for uses such as mass surveillance of Americans or fully autonomous weapons, and this controversy illustrates the tension in the relationship between Anthropic and the government.

Clark also addressed the situation following reports that Trump administration officials recommended Mythos testing to JPMorgan Chase, Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley, among others. On the labor market, he said only weak signs have appeared so far in early graduate employment in some industries, and he advised that college majors should focus on developing the ability to synthesize across multiple fields and think analytically.