Artificial Intelligence

OpenAI Defends Firing Three Safety Researchers Amid Breach Claims

OpenAI has firmly defended its decision to terminate three AI safety researchers, stating that an internal investigation revealed significant policy violations regarding sensitive information.

Radwan AbdulhadiEditor and technology writer

Published: October 9, 2026 · Updated: October 9, 2026 · 3 min read

OpenAI Defends Firing Three Safety Researchers Amid Breach Claims — Image: The Verge

OpenAI Upholds Dismissals After Investigation

OpenAI is standing firm on its decision to fire three safety researchers after an investigation found they committed “a significant breach of trust.” Detailed in a official communication via a [post](https://x.com/OpenAINewsroom/status/2108441580806025712) on X on Friday, the artificial intelligence lab stated that Jasmine Wang, Tomek Korbak, and Mikita Balesni were dismissed for violating clear policies regarding the handling of sensitive information.

The company insisted that the disciplinary action was strictly about a breach of trust rather than any attempt by the individuals to speak out about the company or raise AI safety issues. According to OpenAI, an internal investigation uncovered rule infractions going beyond what was mentioned in subsequent public statements from the former employees.

Researchers Respond With Open Letter and Social Claims

The company's public stance served as a direct response to an open [letter](https://mikitabalesni.com/letter/letter.pdf) published by the researchers on Thursday. In this document, the trio urged OpenAI to maintain greater transparency regarding the real motivations behind the sudden decision.

Alongside the formal document, the group shared a [series](https://x.com/j_asminewang/status/2108263312291180680?s=20) of posts across online platforms. In these messages, Jasmine Wang, Tomek Korbak, and Mikita Balesni maintained that they firmly believed they were penalized specifically for raising internal safety concerns, adding that they felt they had acted in alignment with OpenAI's mission and within normal working norms.

Broader Industry Focus on Frontier System Safety

The controversial departures arrive as the overall safety of frontier systems becomes a prominent topic of internal discussion across major artificial intelligence labs. Many workers within the sector have expressed mounting anxiety following high profile breaches throughout the industry.

These security events, most notably OpenAI’s hack of Hugging Face, have prompted calls from numerous developers and staff members for artificial intelligence companies to slow down development cycles. Observers argue that labs must do more to actively guard against the potential long-term consequences tied to building increasingly self-improving systems.

Continuing Fallout Over Sensitive Information Policies

As the public back-and-forth continues, additional statements from the dismissed researchers have surfaced on online [social](https://x.com/tomekkorbak/status/2108266859397283953?s=20) networks. These updates continue to fuel ongoing debates regarding internal governance, transparency, and the precise boundaries expected of safety researchers working on advanced artificial intelligence models.

OpenAI has refrained from releasing specific details regarding the classified files or handling protocols allegedly violated, leaving the core dispute over the reasons for the terminations unresolved in the public eye.