OpenAI Employee Dismissals
OpenAI Fires Three Employees for Mishandling Sensitive Company Information and Rules
Anonymous
Key Points
1
Before the dismissals, 27-year-old researcher Jacob Coxon resigned from Anthropic and warned that leading AI labs, including OpenAI, were “gambling with our lives” by racing to build more powerful models.
2
The named researchers had publicly voiced strong concerns about AI risks: Mikita Balesni posted that he believed AI had more than a 10% chance of killing all humans, while Jasmine Wang warned of the dangers of accelerating toward recursive self-improvement.
3
The dismissals were reported amid a White House meeting on AI involving leaders from OpenAI, Anthropic, Nvidia, SpaceX, Meta and Google. President Donald Trump afterward shared what he called a “morally binding” agreement to protect against AI risks, which critics said left AI companies to regulate themselves.
OpenAI fired three employees—safety researchers Jasmine Wang and Tomek Korbak, plus research program manager Mikita Balesni—for mishandling sensitive company information and violating internal policies Fox Business. The researchers shared material with an outside organization that evaluates AI models, though OpenAI has not disclosed which group received the information or what was shared Crypto Briefing.
The dismissals come as regulators scrutinize AI risks and safety practices across the industry. All three fired employees had publicly warned of dangers from rapid AI development—including Balesni's claim that AI poses more than a 10% chance of human extinction Crypto Briefing.
Three Safety Voices Removed Amid AI Risk Warnings
The three dismissed researchers worked on OpenAI's Safety and Alignment units SEDaily. Jasmine Wang had warned of dangers from accelerating toward recursive self-improvement in AI systems. Mikita Balesni publicly stated he believed AI had more than a 10% chance of killing all humans. Tomek Korbak's specific safety concerns were not detailed in reports Crypto Briefing.
The timing is notable. Just before the firings, Jacob Coxon, a 27-year-old researcher at competitor Anthropic, resigned and accused leading AI labs—including OpenAI—of "gambling with our lives" by racing to build more powerful models without adequate safety guardrails The Online Citizen.
OpenAI's Vague Stance on What Was Actually Shared
OpenAI has not identified the external organization that received the sensitive information or specified what material was involved Crypto Briefing. The company's internal investigation found violations of company policy, but details remain unclear. An outside AI safety organization evaluated the company's models, according to Fox Business, but the exact recipient remains unnamed.
This opacity fuels questions about whether researchers were disciplined for genuine security breaches or for raising uncomfortable truths about AI risks. The lack of transparency contrasts with OpenAI's stated commitment to external oversight of its safety practices The Online Citizen.
White House AI Agreement Leaves Industry Self-Regulating
The firings coincided with a White House meeting attended by leaders from OpenAI, Anthropic, Nvidia, SpaceX, Meta, and Google. President Donald Trump announced what he called a "morally binding" agreement to protect against AI risks TipRanks. Critics immediately flagged that the agreement left AI companies largely responsible for their own safety compliance.
The combination of dismissed safety researchers and industry-friendly regulation raises concern among AI safety advocates. If OpenAI fires employees for sharing safety concerns with external groups, while facing minimal government oversight, the incentive to raise risks internally shrinks considerably The Online Citizen.
Sentiment
Mostly NegativeNEGATIVENEUTRALPOSITIVE
Publishers
22
Articles
40
Reach
62
Topics
A
AnonymousAuthor
