OpenAI Researcher Quits, Flames Company for Axing Team Working to Prevent Superintelligent AI From Turning Against Humankind
OpenAI Shut
It might sound like a joke, but OpenAI has dissolved the team responsible for making sure advanced AI doesn't turn against humankind.
Yes, you read that right. The objective of the team, formed just last summer, was to "steer and control AI systems much smarter than us."
"To solve this problem within four years, we’re starting a new team, co-led by Ilya Sutskever and Jan Leike, and dedicating 20 percent of the compute we’ve secured to date to this effort," the company wrote at the time. "We’re looking for excellent ML researchers and engineers to join us."
If those two names sound familiar, it's because Sutskever departed the company under a dark cloud this week, prompting Leike to quit in disgust.
And now the other shoe has dropped: as first reported by Wired, the entire team has now been dissolved.
Terminator Prequel
Sutskever, who was intimately involved with last year's plot to out OpenAI CEO Sam Altman, has remained largely mum this week. But Leike has been publicly sounding off.
"I joined because I thought OpenAI would be the best place in the world to do this research," he wrote on X-formerly-Twitter. "However, I have been disagreeing with OpenAI leadership about the company's core priorities for quite some time, until we finally reached a breaking point."
Among his gripes: that the company wasn't living up to its promises to dedicate technical resources to the effort.
"Over the past few months my team has been sailing against the wind," he continued. "Sometimes we were struggling for compute and it was getting harder and harder to get this crucial research done."
But his objections also sound more existential than just company politics.
"Building smarter-than-human machines is an inherently dangerous endeavor," Leike wrote. "OpenAI is shouldering an enormous responsibility on behalf of all of humanity."
"But over the past years, safety culture and processes have taken a backseat to shiny products," he alleged.
OpenAI, for its part, has been busy doing exactly that: this week, it showed off a new version of ChatGPT that can respond to live video through a user's smartphone camera in an emotionally inflected voice that Altman compared to the 2013 romantic tragedy "Her."
"It’s a process of trust collapsing bit by bit, like dominoes falling one by one," one OpenAI employee told Vox.
More on OpenAI: The Person Who Was in Charge of OpenAI's $175 Million Fund Appears to Be Fake
