Miles Brundage, a longtime policy researcher at OpenAI and senior advisor to the corporate's AGI readiness team, has left.
In one post on X on Wednesday and in an essay in his newsletterBrundage said he believes he could have more impact as a researcher and advocate within the nonprofit sector, where he could have “more opportunities to publish freely.”
“One reason this decision was difficult is that now greater than ever, working at OpenAI is an incredibly powerful opportunity,” said Brundage. “OpenAI needs individuals who care concerning the mission and are committed to maintaining a culture of rigorous decision-making around development and deployment (including internal deployment, which can turn into increasingly vital over time).”
With Brundage's departure, OpenAI's economic research division, until recently a sub-team of AGI Readiness, will probably be under OpenAI's latest chief economist, Ronnie Chatterji. The remainder of the AGI standby team — which is disbanding — will probably be distributed amongst other OpenAI departments, Brundage says. Joshua Achiam, director of mission alignment, will tackle some AGI readiness projects.
An OpenAI spokesperson told TechCrunch that the corporate “fully supports” Brundage’s decision to pursue its policy research outside of industry and is “deeply grateful” for his contributions.
“Brundage’s plan to completely concentrate on independent research on AI policy gives him the chance to affect more broadly, and we stay up for learning from his work and tracking its impact,” the spokesperson said in a press release. “We are confident that in his latest role, Miles will proceed to lift the bar for the standard of policymaking in industry and government.”
The spokesman didn’t say who will take over Brundage's duties.
Brundage joined OpenAI in 2018, where he began as a research scientist and later became the corporate's head of policy research. Before OpenAI, Brundage was a research fellow on the Future of Humanity Institute on the University of Oxford.
On the AGI on-call team, Brundage placed particular emphasis on the responsible use of voice generation systems similar to ChatGPT. He led other initiatives elsewhere, including OpenAI's external red teaming program and its first “system maps” reports, which document the capabilities and limitations of AI models.
In recent years, OpenAI has been accused by several former employees – and board members – of prioritizing business products on the expense of AI safety. In his post on
“Some people have told me that they’re sad to see me go and that they appreciate that I used to be often willing to lift concerns or questions while I used to be here…OpenAI has many difficult decisions ahead of it and won’t make these “Making right decisions after we succumb to groupthink,” he wrote.
OpenAI has fired high-profile executives in recent weeks, marking the peak of disagreements over the corporate's direction. CTO Mira Murati, Chief Research Officer Bob McGrew and Research VP Barret Zoph announced their resignations at the top of September. Well-known research scientist Andrej Karpathy left OpenAI in February; Months later, OpenAI co-founder and former chief scientist Ilya Sutskever resigned, together with former security chief Jan Leike. In August, co-founder John Schulman said he was leaving OpenAI. And Greg Brockman, the corporate's president, is on an prolonged vacation.
It was a moderately unflattering day for OpenAI.
On Wednesday morning, the corporate was the topic of a New York Times report profile by former OpenAI researcher Suchir Balaji, who said he left the corporate because he now not desired to contribute to technologies that he believed would do more harm than good to society. Balaji also accused OpenAI of violating copyright law by training its models on IP-protected data without permission – an allegation that others have made in school motion lawsuits against the organization.
