CEO of OpenAI and artificial intelligence poster boy Sam Altman has advocated for legislation to guarantee that mankind will eventually develop an AI that will benefit humanity.
The CEO’s actions, however, are in conflict with his public statements, according to a renowned AI expert, and the current trajectory of AI is incorrect.
Gary Marcus, the founder of the machine learning company Geometric Intelligence and the former head of Uber’s AI lab, questioned the sincerity of the CEO’s calls for regulations and claimed that Altman has repeatedly misled the public about his financial stake in OpenAI.
After testifying before Congress about the technology in May 2023, Marcus expressed hope for Altman and said in the editorial that the OpenAI CEO “seemed genuine” and seemed to share the same concerns about AI regulation.
“We both strongly advocated for regulation of AI,” Marcus wrote. “Little by little, though, he realized that I, the Senate, and ultimately the American people, had probably been played.”
Citing the CEO’s indirect financial connections to the company through his interest in Y Combinator, a startup incubator Altman once founded, and Rain AI, a semiconductor startup that struck a $51 million agreement with the AI company, the AI expert contested Altman’s claim that he had no equity in OpenAI.
In addition, Marcus questioned Altman’s dedication to security, noting that while OpenAI claims to support regulation, “the reality is far more complicated.”
He referenced an article from the 2023 issue of Time magazine that described how OpenAI attempted to undercut the EU’s AI Act by eliminating a section designating OpenAI’s ChatGPT as “high risk,” which would have made the firm vulnerable to more stringent regulations.
Marcus also brought up concerns about OpenAI’s lack of transparency, citing instances where staff members were required to sign restrictive nondisclosures agreements and past coworkers accused Altman of misleading the board.
Altman presumably does not wish to live a life of regret and notoriety. However, Marcus added, his lobbyists continue to fight behind closed doors for less regulation or none at all. He went on to say that in an attempt by businesses to catch up to OpenAI, AI is going in the wrong direction beyond Altman and regrettably, a lot of other AI startups appear to be following Altman’s lead of overhype and short cuts.
Additionally, the AI expert contended that ChatGPT and other generative AI tools are “unlikely ever to be safe” and won’t be helpful in fields like health or combating climate change.
According to Marcus, generative AI technologies are inherently opaque and resistive, making them so-called “black boxes” that humans can never fully control.
Despite this, he believes we shouldn’t give up on AI. Improving AI has the potential to change the world significantly, he stated, especially in the fields of medicine, material science, climate research, and so forth. While generative AI is unlikely to work, other, as-yet-undeveloped forms of AI might.
