Published: Saturday, May 18, 2020
Former OpenAI executive who quit the company this week claims that product safety “has taken a backseat” to the company’s “shiny products”.
Former OpenAI leader, who resigned earlier this week from the company, said on Friday that safety had “taken the backseat” to the “shiny products” at the influential artificial-intelligence company.
Jan Leike who led OpenAI’s team of “Superalignment”, along with a co-founder, resigned from the company this week. He wrote on the social media platform X in a series posts that he had joined the San Francisco company because he believed it was the best place for AI research.
Leike’s last day, Thursday, was a “breaking point” for Leike.
Leike, an AI researcher by trade, believes that there should be a greater focus on preparing the next generation AI models. This includes things like safety, and analyzing the social impacts of such technology. He said that building “smarter than human machines” is an inherently risky endeavor and that the company has an “enormous responsibility for all humanity.”
Leike wrote, “OpenAI should become a company that puts safety first. AGI is a futuristic vision for machines to be as intelligent as humans, or at least capable of doing many things just as well.
Open AI CEO Sam Altman replied to Leike’s posts by saying that he “super appreciated” Leike’s contributions to Open AI and was “very sorry to see him go.”
Altman agreed that Leike was “right; we have more work to do. We are committed to it.” Altman promised to write a larger post about the topic in the next few days.
The company confirmed on Friday that Leike’s Superalignment Team, launched last year with the goal of focusing on AI risks and integrating its members into all research efforts, has been disbanded.
Leike’s resignation was announced after OpenAI cofounder and chief scientific officer Ilya Suksever announced Tuesday that he would be leaving the company following nearly a decade. Sutskever voted with four other board members to remove Altman last fall, only to have him quickly reinstated. Sutskever was the one who informed Altman in November last year that he would be fired. He later regretted his actions.
Sutskever stated that he was working on a project that is meaningful to him, without providing any additional details. Jakub Pahocki will replace him as chief scientist. Altman said Pachocki was “also easily one the greatest minds in our generation.” He added that he’s “very confident” he would lead us “to make rapid and secure progress towards our mission to ensure AGI benefits all.”
OpenAI showcased its latest artificial intelligence model on Monday. It can mimic human cadences and even detect moods.
—-
OpenAI and The Associated Press have a technology and licensing agreement which allows OpenAI to access a part of the AP text archives.
Source: ABC News