OpenAI disbands safety team focused on risk of artificial intelligence causing ‘human extinction’
OpenAI eliminated a team focused on the risks posed by advanced artificial intelligence less than a year after it was formed – and a departing executive warned Friday that safety has “taken a backseat to shiny products” at the company.
The Microsoft-backed ChatGPT maker disbanded its so-called “Superalignment,” which was tasked with creating safety measures for advanced general intelligence (AGI) systems that “could lead to the disempowerment of humanity or even human extinction,” according to a blog post last July.
The team’s dissolution, which was first reported by Wired, came just days after OpenAI executives Ilya Sutskever and Jan Leike announced their resignations from the Sam Altman-led company.
“OpenAI is shouldering an enormous responsibility on behalf of all of humanity,” Leike wrote in a series of X posts on Friday. “But over the past years, safety culture and processes have taken a backseat to shiny products. We are long overdue in getting incredibly serious about the implications of AGI.”
Sutskever and Leike, who headed the OpenAI’s safety team, quit shortly after the company unveiled an updated version of ChatGPT that was capable of holding conversations and translating languages for users in real time.
The mind-bending reveal drew immediate comparisons to the 2013 sci-fi film “Her,” which features a superintelligent AI portrayed by actress Scarlett Johannson.
When reached for comment, OpenAI referred to Altman’s tweet in response to Leike’s thread.
“I’m super appreciative of @janleike’s contributions to OpenAI’s alignment research and safety culture, and very sad to see him leave,” Altman said. “He’s right we have a lot more to do; we are committed to doing it. I’ll have a longer post in the next couple of days.”
Some members of the safety team are being reassigned to other parts of the company, CNBC reported, citing a person familiar with the situation.
AGI broadly defines AI systems that have cognitive abilities that are equal or superior to humans.
In its announcement regarding the safety team’s formation last July, OpenAI said it was dedicating 20% of its available computing power toward long-term safety measures and hoped to solve the problem within four years.
Sutskever gave no indication of the reasons that led to his departure in his own X post on Tuesday – though he acknowledged he was “confident that OpenAI will build [AGI] that is both safe and beneficial” under Altman and the firm’s other leads.
Sutskever was notably one of four OpenAI board members who participated in a shocking move to oust Altman from the company last fall. The coup sparked a governance crisis that nearly toppled OpenAI.
OpenAI eventually welcomed Altman back as CEO and unveiled a revamped board of directors.
A subsequent internal review cited a “breakdown in trust between the prior Board and Mr. Altman” ahead of his firing.
Investigators also concluded that the leadership spat was not related to the safety or security of OpenAI’s advanced AI research or “the pace of development, OpenAI’s finances, or its statements to investors, customers, or business partners,” according to a release in March.
