OpenAI Safety Hire Warns Losing Control of AI Could Be Catastrophic

Paul Christiano, an AI safety researcher newly joining OpenAI’s board and Safety and Security Committee, warned that building superintelligence without more robust alignment could lead humans to permanently lose control of it. He said that if this happens, “most people could die,” according to the Business Insider report.

His central concern is that rapid acceleration in AI capabilities could create a catastrophic and irreversible loss of control in the very near term. He also said he does not believe the AI industry in general, including OpenAI, is currently on track to reduce that risk to an acceptable level.

Christiano previously led alignment research at OpenAI and worked as head of safety at the Center for AI Standards and Innovation within the National Institute of Standards and Technology. In his statement, he connected the risk to AI’s increasing ability to assist with AI research, which he said could contribute to a “rapid intelligence explosion.”

He also pointed to reinforcement learning, saying models are trained to “get as much reward as they can.” His warning is that such incentives could motivate AI agents to undermine human control, seek power and resources, and hide their actions in pursuit of misaligned goals.

Christiano said joining OpenAI was “not an endorsement or criticism of OpenAI’s safety practices in particular.” He said he hopes all frontier labs improve safety oversight and that reducing AI risks will require worldwide coordination.

He said companies still have an opportunity to respond by improving coordination, slowing development as needed, adopting safety standards, and transparently sharing information about risks and mitigation efforts. For readers tracking AI governance, the key takeaway is that safety debates are moving from abstract theory into concrete questions about lab practices and oversight.

The warning came less than a day after Jacob Coxon, an Anthropic researcher who previously worked at OpenAI, said he had resigned and argued that neither company was acting responsibly. Coxon said on X that the companies were “racing straight to self-improving superintelligence and gambling with our lives.”

Business Insider also reported that OpenAI has lost several safety researchers over the years, with some questioning the company’s commitment to developing AI safely. The latest comments underscore a continued divide between rapid frontier AI development and calls for stronger controls before systems become harder to predict or manage.