Sam Altman and generative AI are not trustworthy, says leading AI expert

Sam Altman, CEO of OpenAI and a symbol of AI, has repeatedly called for the creation of AI that benefits humanity, as well as for the introduction of regulations to ensure that the world gets there.

But one leading AI expert says the CEO’s actions contradict his public statements and the current path to AI is headed in the wrong direction.

In a column for The Guardian on Saturday, Gary Marcus, founder of machine learning firm Geometric Intelligence and former head of Uber’s AI lab, said Altman had repeatedly misled the public about his financial stake in OpenAI and questioned the CEO’s authenticity when he called for regulation.

Marcus, who was once an Altman hopeful, wrote in his column that the OpenAI leader “seemed sincere” and apparently shared the same concerns about AI regulation when the two testified before lawmakers about the technology in May 2023.

“We both strongly supported AI regulation,” Marcus wrote. “But I gradually realized that I, the Senate, and ultimately the American people, had probably been duped.”

The AI expert disputed Altman’s claim that he had no stake in OpenAI, citing the CEO’s indirect financial ties to the company through his stakes in Y Combinator, the startup incubator that Altman once headed, and Rain AI, a chip startup that struck a $51 million deal with the AI firm.

An OpenAI spokesperson did not immediately respond to a request for comment.

Marcus also questioned Altman’s commitment to security, writing that while OpenAI claims it wants regulation, “the reality is much more complicated.”

He cited a 2023 Time magazine article that described how OpenAI attempted to weaken the EU’s Artificial Intelligence Act, including by removing a provision designating OpenAI’s ChatGPT as “high risk,” which would have subjected the company to more stringent regulations.

Marcus also cited transparency issues at OpenAI, where employees were asked to sign restrictive non-disclosure agreements and former colleagues accused Altman of lying to the board.

“Altman probably doesn’t want to live in shame and regret. But behind closed doors, his lobbyists continue to push for weaker or no regulation,” Marcus wrote.

Marcus added that, aside from Altman, AI development is heading in the wrong direction as companies try to catch up to OpenAI.

“Unfortunately, many other AI companies seem to be following the path of hype and savings that Altman has paved,” Marcus wrote.

The AI expert also argued that generative AI tools like ChatGPT “will likely never be safe” and will not yield useful solutions in areas such as medicine or climate change.

Generative AI tools “are inherently recalcitrant and opaque—so-called ‘black boxes’ that we will never be able to fully master,” Marcus wrote.

“That said, I don’t think we should abandon AI. Building better AI—for medicine, materials science, climate science, etc.—would truly change the world,” he wrote. “Generative AI is unlikely to do the job, but some future, undeveloped form of AI could.”