OpenAI company faces internal turbulence after the departure of Ilya Sutskever and Jan Leike, and reviews its priorities in Artificial Intelligence research.
OpenAI recently confirmed the disbandment of its Superalignment Team, a research group dedicated to developing strategies to control and guide superintelligent artificial intelligence. This development comes after a number of key departures, including that of Ilya Sutskever, co-founder and chief scientist of OpenAI, who he was one of the co-leaders of the team.
The Superalignment Team had been announced by OpenAI in July last year with the aim of preparing for the arrival of superintelligent AI, capable of surpassing and potentially overpowering their creators. The formation of the team had been seen as a significant step towards managing the long-term risks associated with artificial intelligence. OpenAI had allocated 20 percent of its computing power to this group, signaling the importance of its mission.
Leadership changes at OpenAI
The team’s recent dissolution was preceded by the departure of several prominent researchers. The news that Ilya Sutskever reportedly left OpenAI caused a sensation, since in addition to being one of the founders, he was one of the four board members who they fired CEO Sam Altman last November. Altman was reinstated after five chaotic days, thanks to a staff revolt and a deal that saw Sutskever and two other directors leave the board.
A few hours after Sutskever’s departure was announced, Jan Leike, the other co-leader of the Superalignment Team, also announced his resignation. In a post on X (formerly Twitter), Leike explained that his decision was caused by disagreements over company priorities and the amount of resources allocated to his team. “For the past few months, my team has been sailing against the wind. At times we struggled for computing power and it was increasingly difficult to carry out this crucial research,” Leike wrote.
And now?
Research into the risks associated with the most powerful AI models it will now be led by John Schulman, who will coordinate the team responsible for fine-tuning the AI models after training. OpenAI declined to comment on the departures of Sutskever or other members of the Superalignment Team, or the future of his work on the long-term risks of AI.
Regardless, OpenAI continues to make significant progress in AI development. Recently, the company presented a new multimodal AI model called GPT-4ocalled GPT-4o, which allows ChatGPT to see the world and converse in a more natural and human way. This model raises new ethical questions regarding privacy, emotional manipulation and cybersecurity risks, underlining the need for adequate regulation.
#OpenAI #abandonment #Jan #Leike #Superalignment #Team #leaves
