The co-founder of Anthropic confirms that the company explained the Trump administration on Mythos

Jack Clark, one of Anthropic’s co-founders who also serves as the Head of Public Benefit for Anthropic PBC, confirmed that the AI ​​company told the Trump administration about its new version of Mythos.

model, announced last weekit is so dangerous that it is not released to the public, mainly because of its cyber security capabilities.

In the discussion to Semaphore World Economy summit this week, Clark explained why the company is still engaged with the US government while at the same time suing them.

The month of March, Anthropic filed a lawsuit against Trump’s Department of Defense (DOD) after the agency called the company a security risk. Anthropic has argued with the Pentagon over whether the military should have access to Anthropic’s AI systems for crimes that include mass surveillance of Americans and autonomous weapons. (OpenAI ended win the deal instead.)

At the meeting, Mr. Clark downplayed the government’s claims that his business was in danger of being sold, saying that it was a “low level argument” and that Anthropic did not want it to be a distraction because the company was concerned about national security.

“Our position is that the government must know about this, and we must find new ways to make the government cooperate with the private sector that is creating things that change the economy, but they will have things that hit National Security, equities, etc.,” said Clark. “Absolutely, we talked to them about the Mythos, and we will talk to them about the next versions as well.”

His confirmation comes later reports last week that the Trump administration was encouraging banks to test Mythosincluding JPMorgan Chase, Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley.

Clark also covered some of AI’s impact on society during the interview, including things like unemployment and higher education.

Earlier, Anthropic CEO Dario Amodei warned that the progress of AI could lead to unemployment to Depression-era numbersbut Clark disagrees a little. He explained in the interview that Amodei believes that AI will gain more power than people expect very quickly, so he is using this as his foundation.

Clark, who leads the group of economists at Anthropic, said the company is only seeing “some potential weakness in graduate employment” in select industries. He added that Anthropic is prepared if there is a major change in service, however.

When pushed to say what college students today should pursue or avoid, given the challenges of AI, Clark could only say that the majors are those that “relate across disciplines and think critically about it.”

“That’s why AI allows us to do it is that it gives you access to multiple subject matter experts in different fields,” Clark said. “But the most important thing is to know the right questions to ask and to have an idea of ​​what will interest you when you meet different ideas from different disciplines.”