OpenAI Safety Committee Gains Independence as Altman Steps Down
Sam Altman leaves OpenAI's Safety and Security Committee, which transitions to an independent board oversight group led by Zico Kolter. The reshuffle follows intense scrutiny from U.S. senators and high-profile departures from the company's safety teams.
Sam Altman steps down from OpenAI's Safety and Security Committee as the organization transitions the group into an independent board oversight body. Carnegie Mellon professor Zico Kolter now chairs the committee, which includes existing board members Adam D’Angelo, Paul Nakasone, and Nicole Seligman. The newly independent group recently completes a safety review of OpenAI's latest AI model, o1, and retains the authority to delay product releases until safety concerns are fully addressed.
This leadership change follows increased pressure from five U.S. senators who question OpenAI's safety policies earlier this summer. Nearly half of the staff dedicated to long-term AI risks have departed the company, and former researchers accuse Altman of favoring corporate interests over genuine AI regulation. OpenAI significantly increases its federal lobbying budget to $800,000 for the first half of 2024, a sharp increase from $260,000 for all of 2023.
Despite Altman's removal from the committee, critics doubt the group will make decisions that negatively impact OpenAI's commercial roadmap. The company states the committee will continue to receive regular technical assessments and post-release monitoring reports to establish an integrated safety framework. Observers note that OpenAI's definition of "valid criticisms" remains subjective, leaving uncertainty about the committee's true independence and willingness to challenge the company's aggressive product launch strategy.