Why Anthropic’s Massive Code Leak Is Now a National Security Concern in D.C.

A lawmaker is demanding answers after Anthropic accidentally exposed internal code, warning the leak could erode the U.S. AI advantage.

BY LEILA SHERIDAN, NEWS WRITER

Illustration: Inc; Photo: Anthropic, Getty Images

Earlier this week, Anthropic accidentally released more than 500,000 lines of code tied to its Claude Code system, offering an unusually detailed look into how one of the fastest-growing AI tools operates. Now, the incident is raising new concerns in Washington about national security and the role of AI in defense. 

Representative Josh Gottheimer, a New Jersey Democrat and leading House voice on AI and cybersecurity, wrote to Anthropic CEO Dario Amodei on Thursday warning of potential risks tied to the leak, according to a letter obtained by Axios. His message reflects mounting pressure from lawmakers as AI systems become increasingly embedded in U.S. defense and intelligence operations.

Leaks Raise Lawmaker Concerns

The letter follows what Gottheimer described as the second time in just over a year that source material powering Claude has leaked publicly. “Claude is a critical part of our national security operations. If it is replicated, we sacrifice the competitive edge we have worked so diligently to maintain in all facets of our national security,” he wrote, according to Axios.

At the same time, Gottheimer made clear he is at odds with the Trump administration’s approach to Anthropic. He voiced support for the company amid its ongoing tensions with the Pentagon, writing, “I was deeply concerned with the Trump Administration’s decision to block Claude from all government contracts, and how that decision could significantly set us back in our AI race against China.”

How Anthropic's Claude AI Became a Co-Founder

That dual stance—defending Anthropic’s role in national security while pressing it on internal safeguards—underscores the increasingly complex position AI companies now occupy.

Gottheimer’s letter also raises concerns about whether Anthropic has weakened certain safety protocols. He pointed to a Chinese Communist Party-backed group that hacked Claude last year and warned that the company’s upcoming model, Mythos, could be exploited to carry out more advanced cyberattacks. “Anthropic needs to do everything it can to prevent CCP-backed companies like DeepSeek from conducting distillation campaigns,” he wrote.

Anthropic Tries to Assuage Concerns

Anthropic, for its part, has downplayed the severity of the incident. “A Claude Code release included some internal source code. No sensitive customer data or credentials were involved or exposed,” a company spokesperson told Axios following the leak Tuesday. “This was a release packaging issue caused by human error, not a security breach. We’re rolling out measures to prevent this from happening again.”

© Inc.com