In a stunning turn of events, key figures from OpenAI’s AI safety team have resigned, raising critical questions about the future of artificial intelligence oversight. Jan Leike, a top machine learning researcher, confirmed his departure with a succinct “I resigned,” closely following the exit of Ilya Sutskever, a cofounder of OpenAI and the chief scientist of the company. These departures mark a significant shift within the organization that has been at the forefront of developing cutting-edge AI technologies.
The resignations of Leike and Sutskever come amid a broader exodus from OpenAI’s safety team, which has experienced a spate of high-profile exits in recent months. This team, known for its efforts to align artificial intelligence systems with human interests and prevent them from turning rogue, was co-led by the pair. In their roles, they were tasked with ensuring that the burgeoning capabilities of AI remain beneficial and under control—a mandate of paramount importance given the potential risks associated with superintelligent systems.
On the heels of these resignations, Sam Altman, CEO of OpenAI, expressed his sentiments on the social platform X, acknowledging Sutskever’s profound influence on the field and the company, “Ilya is easily one of the greatest minds of our generation, a guiding light of our field, and a dear friend.”
The Superalignment team’s mission, as detailed in a July 5, 2023, post on OpenAI’s website, underscored the urgency of developing controls for superintelligent AI systems. “We need scientific and technical breakthroughs to steer and control AI systems much smarter than us,” the company stated, emphasizing the need for innovative solutions beyond current techniques such as reinforcement learning from human feedback.
Leike, prior to joining OpenAI, had contributed significantly to Google’s DeepMind and harbored grand aspirations for solving the alignment problem—the issue of ensuring machines act in accordance with human intentions. On the “80,000 Hours” podcast in August 2023, Leike expressed optimism about the possibility of making strides in AI safety, “It’s like we have this hard problem that we’ve been talking about for years and years and years, and now we have a real shot at actually solving it.”
In March 2022, Leike also laid out on his Substack a strategy to achieve alignment. “Maybe a once-and-for-all solution to the alignment problem is located in the space of problems humans can solve. But maybe not,” Leike wrote, “by trying to solve the whole problem, we might be trying to get something that isn’t within our reach. Instead, we can pursue a less ambitious goal that can still ultimately lead us to a solution, a minimal viable product (MVP) for alignment: Building a sufficiently aligned AI system that accelerates alignment research to align more capable AI systems.”
Relevant articles:
– It was their job to make sure humans are safe from OpenAI’s superintelligence. They just quit., businessinsider.com, 05/17/2024
– Sam Altman gracefully thanked his OpenAI cofounder who quit. Then another exec quit hours later., Business Insider, 05/15/2024
– OpenAI Staffers Responsible for Safety Are Jumping Ship, Gizmodo, 05/16/2024
– Ilya Sutskever, Co-Founder and Chief Scientist, Leaves OpenAI, TIME, 05/15/2024
– OpenAI’s Chief AI Wizard, Ilya Sutskever, Is Leaving the Company, WIRED, 05/15/2024
Glad you enjoyed above story, be sure to follow TrendyDigests on Microsoft Start.