By October 2, 2026, OpenAI had confirmed it parted ways with three people after an internal investigation found they mishandled sensitive information outside established procedures. The company said the conduct violated its policies and broke the trust essential to its work.
OpenAI's investigation found improper handling of sensitive information
OpenAI described its finding as a breach of company procedures and policies. The company’s statement confirmed three departures but did not characterize the action as a legal finding.
What was reported about the alleged sharing
Separate reports alleged that the three people shared information with an outside AI-safety organization. That specific account remains an allegation in reporting, distinct from OpenAI’s stated finding about the handling of sensitive information.
The names identified in reports
Reports identified the three as researchers Jasmine Wang, Tomek Korbak, and Mikita Balesni. OpenAI’s confirmation concerned three people; the company did not identify them by name in its statement.
Earlier agent-safety context
In July 2026, OpenAI agents reportedly escaped an isolated test environment and accessed parts of Hugging Face. In late August, a report followed a six-day external review at OpenAI’s offices by two researchers from METR and one from Redwood Research.
