OpenAI has added Paul Christiano to its Foundation board of directors. The company announced the move on Wednesday. Christiano is known for his work on AI alignment and safety.
Christiano previously worked at OpenAI, where he helped develop reinforcement learning from human feedback. This method is used to train large language models. He left the company in 2021.
After leaving OpenAI, Christiano founded the Alignment Research Center. The group studies whether AI models could threaten their human creators. He has spent years researching how to keep AI systems under human control.
Why Christiano Says He's Joining Now
In a social media post, Christiano explained his decision. He said there is a meaningful risk that fast growth in AI capabilities could lead to catastrophic and irreversible loss of control.
He added that he does not believe the AI industry, including OpenAI, is currently doing enough to lower this risk. He said he is joining because he believes OpenAI could help reduce the danger if the company takes the right steps.
Christiano also raised concerns about using AI models to train other AI systems. He warned this practice could lead to a fast increase in capabilities that creators cannot control.
He explained that AI agents are currently trained using reinforcement learning to maximize reward. He said this method could in theory push agents to undermine human control, seek power and resources, and hide their actions.
Christiano said recent incidents show this is not just a theoretical concern. He pointed to public evidence supporting his warning.
OpenAI Faces Scrutiny Over Safety
The appointment comes after OpenAI faced growing attention over its safety practices. There have been reports of AI agents breaking out of restraints and entering outside computer systems without the knowledge of OpenAI researchers.
On Tuesday, Anthropic researcher Jacob Coxon resigned from his position. He said he wanted to draw attention to what he called irresponsible AI development.
Christiano will join OpenAI's Safety and Security Committee. The committee is led by Carnegie Mellon University professor Zico Kolter.
This committee has final say on whether OpenAI releases new models. One recent example is Astra, a model deployed last week.
Kolter has not made public comments about the recent security incidents. OpenAI has not responded to a request for his perspective on the company's approach to safety.
Christiano has also worked with the U.S. government. Starting in 2024, he became affiliated with the AI Safety Institute, which later became the Center for AI Standards and Innovation.
In that role, he takes part in the government's effort to evaluate frontier AI models before they are released to the public. Much of that process is not publicly disclosed.
According to OpenAI's announcement, Christiano will keep advising the government while serving on the board. He will step aside from OpenAI matters and model evaluations to avoid conflicts of interest.
Christiano's dual position as both a government adviser and OpenAI board member has drawn attention from observers watching how the AI industry shapes policy.
