OpenAI has cut ties with three researchers from its safety team following alleged breaches of confidentiality, according to a report by The Wall Street Journal. The incident comes amid a series of security lapses involving its AI agents and amid broader concerns over the company's prioritisation of safety.

According to an OpenAI spokesperson, the three individuals were dismissed for mishandling sensitive information, breaking established company policies and damaging trust within the organisation. The report did not provide details on the specific information or the third-party organisation involved.

These departures follow similar incidents in the past, with OpenAI previously firing researchers Leopold Aschenbrenner and Pavel Izmailov in 2024 over alleged leaks. The latest moves come just two days after The New York Times reported that OpenAI executives had reportedly dismissed internal safety concerns, leading to questions about the company's commitment to security.

As OpenAI grapples with a series of containment breaches and public relaunch cancellations, the loss of key safety personnel could have significant implications for the future development and deployment of AI technologies. Whether the remaining team can fill the void and maintain the necessary vigilance remains to be seen.