Three former OpenAI researchers say they were let go after raising safety concerns, sparking a dispute over how the firm handles internal dissent.
Mikita Balesni posted the accusation on X, joined by former colleagues Tomek Korbak and Jasmine Wang. They claim the firings “prioritised safety over near‑term corporate interests.”
OpenAI’s spokesperson, speaking on 2 October, confirmed that an investigation found the trio had “mishandled sensitive information outside established company procedures.” The company reiterated the verdict in a note denouncing any link to safety concerns.
The ex‑employees argue the firings stifle an environment that once encouraged discussion of alignment – the practice of embedding human values into AI systems. Korbak warned that the company will now lose tools for monitoring AI behaviour.
OpenAI has pledged to finalize contracts with third‑party safety assessors and will publish further details soon. The case arrives amid global calls for clearer oversight of AI capabilities and the risk they pose to society.
Implications
If safety advocates feel silenced, the industry may face reputational damage and slowed innovation. The dispute also may prompt regulators to craft stricter guidelines for AI developers, ensuring internal channels remain open for risk concerns.
Technologists like Andrea F., a former Anthropic researcher, have expressed similar fears about a future where paid incentives undermine safety priorities. The debate highlights the tension between commercial ambition and long‑term societal welfare.
