A source confirmed to CNBC on Friday that OpenAI has disbanded its team focused on the long-term risks of artificial intelligence just one year after its formation. Some were reassigned to various other teams within the company.
The Disbandment of OpenAI's Superalignment Team
The disbandment news comes days after OpenAI co-founders Ilya Sutskever and Jan Leike announced their departure from the Microsoft-backed startup.
OpenAI's Superalignment team, announced last year, was dedicated to achieving scientific and technical breakthroughs to steer and control highly advanced AI systems, of which OpenAI had committed 20% of its computing power to this initiative over four years.
OpenAI did not provide a comment, instead directing CNBC to co-founder and CEO Sam Altman's recent post on X. In it, Altman expressed sadness over Leike's departure and acknowledged that the company still had more work to do.
OpenAI Leadership Shakeup
Sutskever and Leike on Tuesday announced their departures from OpenAI on the social media platform X hours apart. However, on Friday, Leike provided more details about his reasons for leaving the startup, expressing his belief that OpenAI would be the best place to conduct research. However, he had disagreed with the leadership about the company's core priorities for some time, leading to a breaking point. As emphasized, there is a need for more focus on security, monitoring, preparedness, safety, and societal impact, but the company was not on a trajectory to address these challenges effectively.
Leike noted the struggles his team faced, including difficulties obtaining computing resources, which hindered crucial research efforts. He added that OpenAI must become a safety-first AGI company. Building smarter-than-human machines poses inherent risks. Leike highlighted that OpenAI carries a significant responsibility for humanity but noted that a focus on flashy products has overshadowed safety culture and processes in recent years.
Altman also expressed sadness about Sutskever's departure on Tuesday and praised him as one of the greatest minds of their generation through his significant contributions to the field. Jakub Pachocki, a research director with OpenAI since 2017, has also been announced as the chief scientist to replace Sutskever.
OpenAI's New AI Model: GPT-4
The news of Sutskever's and Leike's departures, along with the dissolution of the superalignment team, comes shortly after OpenAI launched a new AI model and desktop version of ChatGPT, along with an updated user interface. This represents the company's latest endeavor to broaden the use of its popular chatbot.
During a live-streamed event on Monday, technology chief Mira Murati announced that the update brings the GPT-4 model, including the new GPT-4o variant, to all users of OpenAI, including those on the free tier much faster, with enhanced capabilities in text, video, and audio processing. OpenAI revealed its intention to enable users to video chat with ChatGPT in the future, citing that this development marks a significant advancement in the platform's ease of usability.
