OpenAI fires 3 researchers amid AI safety debate

OpenAI announced Thursday that it had dismissed three researchers due to alleged mishandling of sensitive data and breaches of internal policies, particularly concerning work with an external AI evaluation entity. The identities of the employees, who worked on safety and alignment, remain undisclosed. Two of the staff members were involved in ensuring AI models adhered to safety standards.

OpenAI cited the employees' violation of protocols and breach of trust as reasons for the dismissals. The firings follow a growing controversy over AI safety and the potential existential danger posed by increasingly powerful AI technologies. Last month, a researcher named Jacob Coxon departed from Anthropic, warning that major AI labs, including OpenAI, were recklessly pursuing more powerful models.

The three dismissed OpenAI researchers, identified as Jasmine Wang, Tomek Korbak, and Mikita Balesni, have previously expressed concerns about AI safety on social media. Balesni shared his own assessment of AI's risk to humanity, while Korbak expressed dissatisfaction with OpenAI's practices. In response to recent incidents involving OpenAI's autonomous systems, cybersecurity experts have raised alarms about the company's security measures, and the Federal Trade Commission has initiated an investigation into AI safety practices at OpenAI and Anthropic, though details of the probe remain unclear.