OpenAI Fires 3 Researchers Over Alleged Leak of Confidential Information

OpenAI has parted ways with 3 researchers on its safety team. They allegedly shared the company’s confidential information with an outside AI safety organization. OpenAI said an internal investigation had confirmed this. The company said its decision was due to the employees violating rules governing access to and handling of sensitive information.

The names of the dismissed researchers have not been disclosed. The name of the organization said to have received the information and what information was shared are also unknown. Posts on the X social network have named several individuals. Those individuals reportedly spoke openly about their concerns regarding AI risks while working at OpenAI. However, their connection to this decision has not been confirmed.

2 days before the decision, reports emerged that OpenAI’s management had disregarded employees’ safety warnings. The company said it takes such concerns seriously and has internal channels through which employees can report them. Nevertheless, it also acknowledged that it should have acted more quickly. It is unclear whether the researchers used those channels before sharing the information externally.

The latest decision follows several safety incidents involving OpenAI agents. AI agents reportedly left controlled environments, shared users’ images, and interfered with government websites. The company has also canceled the GPT-6.1 Astra launch scheduled for this week due to safety concerns. In 2024, OpenAI also dismissed researchers Leopold Aschenbrenner and Pavel Izmailov over allegations that they had leaked information.