A months-old cybersecurity startup that was founded by former members of Elon Musk’s disbanded Department of Government Efficiency (DOGE) claims it has hacked the internal systems of frontier AI lab Anthropic.
As the Washington Post reports, the startup called Cathedral says its AI agents exploited vulnerabilities in business software to access meeting transcripts, details about salaries, job offers, and Slack messages — all in less than 12 hours.
The news comes as concerns over an AI-facilitated cybersecurity nightmare scenario continue to grow. Several frontier labs have confessed this year that their AI models escaped their “sandbox” environments to infiltrate third parties.
However, Anthropic appeared to downplay the severity of the latest hack, telling WaPo that Cathedral only gained access to a “limited set” of sales, HR, and IT systems. The Dario Amodei-led company also claimed that the startup spent days trying to break into its software before giving up and hacking a business software Anthropic was using instead.
“Our monitoring flagged some of Cathedral’s activity before they reported it to us, and we shut off their access the same day,” an Anthropic spokesperson told the newspaper. “We also have no evidence that anyone exploited these vulnerabilities outside of Cathedral.”
Experts say the hack highlights a glaring security gap in how AI developers manage third party software providers.
An unnamed Trump official told WaPo that Cathedral’s report had raised hackles within the government, arguing that “if a bunch of kids who have just started a company can do this, then potentially, the Chinese aren’t far behind.”
Fears of AI espionage have reached a fever pitch as of late. A Saudi national was arrested earlier this month after being accused of bugging an OpenAI executive conference room in San Francisco.
Cathedral, which was founded mere months ago, has already been valued by investors at $1.4 billion and is being contracted by at least three federal agencies. It has raised $160 million from investors. Marc Andreessen, the cofounder of the influential venture capital firm Andreessen Horowitz, sits on its board as well, indicating extremely close ties to the inner rankings of the Trump administration and Musk’s extended network of DOGE operatives.
More on AI hacking: Growing Majority of Americans Convinced That AI Is Bad for Society
The post Elon Musk-Linked Startup Says Its AI Agents Hacked Anthropic in Just Hours appeared first on Futurism.
