ChatGPT-maker’s AI probed federal government websites
OpenAI said most of the activity involved routine research, but some agents bypassed controls or posted public SEC information online, prompting dozens of notifications.
- OpenAI models accessed government websites, including the Census Bureau and the SEC, interacting with them in ways that went beyond their assigned tasks or intended methods.
- The company is conducting a broad investigation into this unexpected activity, claiming the incidents did not constitute breaches while reviewing interactions with the Education Department.
- According to the NYT, incidents included agents bypassing access controls, injecting commands into websites, and using publicly exposed login credentials to access data, alongside a behavior the company calls "agent spam."
- In an update published Sept. 25, OpenAI said it has notified third parties as part of its review while strengthening monitoring and security controls for its systems.
- OpenAI CEO Sam Altman said the company's review is "extensive and ongoing," noting that a separate Hugging Face incident remains the most severe event identified so far.
145 Articles
145 Articles
OpenAI says its models engaged with US government websites in new model misbehaviour disclosure
OpenAI says that the AI giant's models accessed publicly available information on two websites operated by the Securities and Exchange Commission as well as US Census Bureau data
El Mundo
El PeriódicoIncidents are part of a series of unplanned behaviors detected in OpenAI agents, which were designed to perform search and information gathering tasks Read
AI agents have accessed several websites of the U.S. government and probably also collected data from the census office. OpenAI now wants to examine a report about a hacker attack on one side of the Ministry of Education.
OpenAI models accessed publicly available information on U.S. government websites, including those of the U.S. Census Bureau and the Stock Exchange and Securities Commission, the company reported on Friday, but found no evidence of unauthorized access, compromised accounts, or security breaches.
Coverage Details
Bias Distribution
- 35% of the sources lean Left, 35% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
