OpenAI Terminates Three Employees Over Improper Handling of Sensitive Data

OpenAI, the company behind ChatGPT, has dismissed three researchers after an internal probe found they allegedly mishandled “sensitive information” and broke company rules—work that included involvement with an outside group that evaluates artificial intelligence models.

The San Francisco-based AI lab did not name the employees in its statement. However, the Wall Street Journal and Bloomberg reported that at least two of those affected worked on safety and alignment, areas at the center of the industry’s growing debate over how to manage powerful AI systems.

“We have parted ways with three individuals,” OpenAI told AFP in a statement.

“Our investigation confirmed that these individuals mishandled sensitive information outside established company procedures, violating our policies and breaking the trust essential to our work.”

The terminations land as the AI world wrestles with high-stakes questions about safety—and whether advanced systems could pose an existential threat to humanity.

That tension sharpened last month when Jacob Coxon, a 27-year-old researcher, resigned from Anthropic and issued a blunt warning that top AI labs—including OpenAI, where he previously worked—were “gaming with our lives” by pushing ahead in a race to build increasingly capable models.

According to the WSJ, the researchers fired by OpenAI are Jasmine Wang, Tomek Korbak, and Mikita Balesni.

In recent weeks, all three had frequently used X to discuss AI safety and related concerns.

“i am at OpenAI and i think AI is >10% likely to kill all humans,” Mr Balesni wrote on 10 September, aligning with similar expressions of alarm from other AI employees in recent weeks.

As questions circulated about why staff at OpenAI and Anthropic were speaking so openly about personal views, Mr Korbak weighed in publicly.

“I’m quite unhappy with much of what OpenAI does. I am very happy that Im allowed to say ‘I’m quite unhappy with much of what OpenAI does’,” Mr Korbak posted on 11 September.

Ms Wang, responding to Mr Coxon’s resignation, highlighted what she described as the risks of pushing too quickly toward recursive self-improvement—software designed to continually teach itself.

“It’s hard to overstate how dangerous speeding towards RSI is,” Ms Wang wrote.

The broader policy backdrop has also shifted this week, with leading US tech firms agreeing to a voluntary self-regulation pledge on AI safety following a White House meeting with President Donald Trump.

Mr Trump described the commitment as “morally binding” and aimed at ensuring adequate safeguards for a rapidly advancing technology.

Executives from Nvidia, Google, Meta, XAI, OpenAI and Anthropic signed the agreement.