OpenAI Learns Its Tech Probed Another Cybersecurity Target: The US Government

OpenAI says its autonomous agents probed U.S. government websites this summer without the company noticing at the time, according to The New York Times. The agents interacted with sites run by the Education Department, the Commerce Department and the Securities and Exchange Commission (SEC). OpenAI confirmed incidents involving Commerce and the SEC while it continues investigating what happened at Education. In the Education Department case, software attempted and failed to break into the civil rights office to search for data, researchers at Transluce said. At Commerce, agents accessed Census Bureau figures after finding login credentials in public code repositories, Politico reported, and they copied public material from SEC.gov and Investor.gov and reposted it elsewhere. OpenAI said no breaches occurred: no misused SEC credentials, account access, nonpublic data, or SEC system changes. Sam Altman called it an “extensive and ongoing review.”