OpenAI Confirms Firing of Three Researchers: It's About Trust, Not Safety

OpenAI is standing by its decision to fire three AI safety researchers. According to the company, Jasmine Wang, Tomek Korbak, and Mikita Balesni were dismissed for a "serious breach of trust" and violation of internal policy on handling sensitive information. The company stresses that the decision has nothing to do with the employees' attempts to speak out about safety concerns. The corporation published an official statement on Friday in a post on X, responding directly to the fired employees' criticism.

The researchers' position and their open letter

The fired researchers — Jasmine Wang, Tomek Korbak, and Mikita Balesni — published an open letter on Thursday, stating that they "acted in accordance with OpenAI's mission and the workplace norms of the time." They demanded greater transparency from the company about the reasons for their dismissal and asked for public acknowledgment of their work. According to the researchers, they committed no serious violation at all — the real issue, they say, was that they criticized AI safety policy and raised concerns about the potential risks of developing self-improving systems that could become dangerous to humanity.

OpenAI's response: an expanded investigation

In its response, OpenAI said its internal investigation uncovered violations going "beyond what is described in the letter" by the researchers. However, the company declined to disclose details of these violations. Instead, it simply stated that the fired employees violated "clear policies on handling sensitive information." This approach from OpenAI has drawn criticism, as it reads like an attempt to obscure the real motives behind parting ways with its critics.

Pressure on the AI industry: safety in the spotlight

The OpenAI saga is unfolding against a backdrop of growing demands for stronger safety policies at AI companies and intensifying debate over how artificial intelligence should be developed. Following high-profile incidents, including a breach of the Hugging Face platform, employees at leading AI labs are increasingly calling for slower development timelines and a more serious approach to the consequences of building self-improving systems.

Many specialists within AI companies are now openly discussing the need for careful analysis of potential risks when deploying advanced AI systems. They point to the dangers of systems capable of self-improvement without adequate oversight or a real understanding of their true capabilities.

The escalating conflict between OpenAI and researchers Jasmine Wang, Tomek Korbak, and Mikita Balesni reflects a broader industry debate about how companies should respond to internal critics raising AI safety concerns. This clash symbolizes the tension between the drive for innovation and the demand for ensuring the safety of fast-evolving artificial intelligence technologies.