WASHINGTON, October 26. /TASS/. Cathedral, a start-up firm founded by former staffers of Elon Musk’s now-defunct Department of Government Effectiveness (DOGE), broke into sensitive programs at US AI giant Anthropic, The Washington Post (WaPo) reported, citing a draft report by Cathedral.

According to it, the fledgling start-up penetrated Antropic’s sensitive internal data, including meeting transcripts and salary details, as well as messaging systems by exploiting vulnerabilities in its business software. Anthropic confirmed that Cathedral had obtained access to a "limited set" of its systems and reported the incident to federal officials, the WaPo said. A source close to Cathedral told the newspaper that the start-up attempted to access Anthropic systems to highlight the security challenges faced by AI developers in the United States.

A White House official interviewed by the WaPo said the Cathedral report had circulated within the government, raising national security concerns. "How little time it took to get into the sensitive systems of a company that is controlling almost nuclear weapons grade material. <…> If a bunch of kids who have just started a company can do this, then potentially, the Chinese aren’t far behind," the newspaper quoted the official as saying.

According to the report, Cathedral disclosed its findings to the artificial intelligence company through the cybersecurity testing program which offers rewards to security researchers for disclosing vulnerabilities.