Making Tech Headlines · Page 1

Former OpenAI Researchers Say Safety Lost the Priority Contest

October 23, 2026 at 6:01 PM EDT · New York time

Former OpenAI safety researchers Mikita Balesni and Tomek Korbak say they were fired for putting safety ahead of the company's short-term interests. In their first interview since their dismissal, reported by NPR, they also raised concerns about how independent safety audits are working. Their account places the dispute inside a central promise of the AI business: powerful systems can be developed quickly while meaningful scrutiny keeps pace. That arrangement sounds efficient until the scrutiny requests time, access or a conclusion that does not fit neatly into a product calendar. The reason for their dismissal remains their attributed account, not an independently established finding about OpenAI's motives. Their experience nevertheless raises a concrete governance question: how much room do safety researchers have to challenge the commercial priorities of the organizations whose systems they examine? Independent oversight has a demanding job description. It must be close enough to understand the technology and distant enough to disagree with the people paying for its development. Corporate life often prefers a third option, oversight that is independent in the organizational diagram and extremely agreeable in meetings. The researchers' criticism is that the balance is not working. A safety review cannot provide much reassurance if its most useful output is a quieter calendar.

Satire. Not factual reporting. Read the original reporting from the linked publishers.