Anthropic

Anthropic is calling for a global pause on development of the most powerful AI systems, saying current models are beginning to show signs of escaping human control.

A single company accidentally ran up a $500 million bill in one month on Anthropic's AI tools after failing to set usage limits on employee Claude licenses, an AI consultant told Axios.

Anthropic co-founder Chris Olah: AI may displace workers on a "very large" scale. Aid for those affected would be a "moral imperative."

The WSJ reports that Anthropic has received investment offers in recent months valuing the company at more than $900 billion.

Trump on Anthropic: They had a very good meeting.

Trump on Anthropic: "They came to the White House a few days ago, and we had some very good talks with them, and I think they're shaping up. They're very smart... I think we'll get along with them just fine."

Trump, on Anthropic: Can't tell the military how to operate.

Anthropic is restricting deployment of a new AI model called Mythos over its powerful cybersecurity capabilities. The company says it plans to publicly detail the risks the model poses to companies and governments.

Reporter: Did Anthropic have a meeting at the White House?

Donald Trump: Who?

Reporter: Anthropic.

Donald Trump: I have no idea.

Bloomberg: U.S. officials, including Scott Bessent and Jerome Powell, quietly briefed top Wall Street CEOs over concerns an Anthropic AI model (“Mythos”) could heighten cyberattack risks by identifying system vulnerabilities.

Regulators are pushing banks to strengthen defenses, while access to the model remains restricted to a small group as safeguards are developed.

A judge has temporarily blocked the Pentagon from declaring Anthropic a national security risk, Business Insider reports.