OpenAI Risk Team Disband Update: Former Leader Claims Company Neglects Safety to Favor 'Shiny Products'

By Giuliano De Leon

May 17 2024, 17:33 PM EDT

OpenAI surprisingly disbanded its entire AI risk team, which was just introduced in July 2023. The former team leader blasted the AI company, accusing it of neglecting safety to prioritize "shiny products."

OpenAI Risk Team Disband Update: Former Leader Claims Company Neglects Safety To Favor 'Shiny Products' — This illustration photograph taken with a macro lens shows The OpenAI company logo reflected in a human eye at a studio in Paris on June 6, 2023. ChatGPT is a conversational artificial intelligence software application developed by OpenAI. JOEL SAGET/AFP via Getty Images

Former OpenAI Risk Team Leader Blasts Company

According to VentureBeat, one of the former leaders of the OpenAI risk team blasted the company. Jan Leike, who announced his departure before the AI safety team's disbandment, posted on X to criticize the tech firm.

"Building smarter-than-human machines is an inherently dangerous endeavor," said Leike via his official X post.

He added that his former company is shouldering a big responsibility on behalf of all humanity. However, Leike claimed that OpenAI neglects safety to favor "shiny products."

Because of this, he warned the world that humanity is already long overdue to take serious action regarding the risks posed by AGI (artificial general intelligence), a type of AI that can perform better than people.

Leike further stated that the world should start preparing for the arrival of AGIs and that OpenAI must become a "safety-first AGI company."

Why Former OpenAI Risk Team Stepped Down Before Disbandment

Aside from accusing OpenAI of neglecting safety and prioritizing new AI products, he explained why he resigned in the first place. Leike said that he joined the company because he thought it would be the best place to do AI research.

Leike was assigned to lead the superalignment team in developing processes and systems that can effectively control superintelligence AI models. Unfortunately, Jan said that he disagreed with OpenAI's leadership.

He said that he believes OpenAI should spend more time on preparing for superalignment, confidentiality, societal impact, monitoring, security, and preparedness for next-gen AI models.

Jan Leike said these issues concern him and make him think they are not on the right path.