OpenAI CEO Sam Altman has been on a real media blitz. But whether he’s doing damage control for his company’s recent scandals or participating in a new Luca Guadagnino biopic, it hasn’t seemed to improve his image. In an interview with Politico, Altman said that “the world should accept some bad things happening for the benefits of this technology,” referring to the large language models being developed by his company and its competitors.
What things does Sam Altman think are bad? He offered a few examples during a two-part interview with Vanity Fair published this week. Altman called it a “bad thing” for kids to be “glued to a touchscreen of some sort scrolling through brain rot short form video,” and said he would limit his own kids’ access to technology. People using LLMs to “create novel pathogens” would also be bad, he asserted. When Vanity Fair editor Mark Guiducci asked him to weigh in on a recent New York Times op-ed by journalist Laura Reiley—whose daughter committed suicide shortly after speaking to ChatGPT—an OpenAI publicist attempted to intervene. Altman eventually answered the question, and equivocated on whether it is good or bad for people experiencing a mental health crisis to turn to an LLM. “This is just like, this is new,” he said. “And the fact that someone in a moment of need that maybe can’t get help or isn’t willing to talk to a therapist can get some help here. Clearly people think there’s a version of this that is good. There are also clearly versions of it that are bad.”
As Altman has made clear in these and other interviews, however, he thinks it would be very bad for governments to regulate his industry in a way that suppresses innovation. He explained to Guiducci that he’s driven by the belief that putting powerful LLMs “in the hands of everyone and giving people a legitimate new superpower relative to I think what any of us could imagine a few years ago will do tremendously more good than bad.” The upshot here is that Altman would like the public to trust OpenAI to judge when the good is outweighing the bad. And although the “goods” of LLM development remain largely speculative—like curing cancer and granting wishes—the bads continue to pile up.
OpenAI faces a growing mountain of legal risks over incidents of its products hacking into government agencies around the world, not to mention providing advice to mass shooters, as Florida’s attorney general’s office has alleged. Nearly every company that’s developing large language models—OpenAI, Nvidia, SpaceX, Microsoft, and Google parent Alphabet—has contracted with a Pentagon that’s been proudly killing fishermen in Latin America, and that continues to wage an illegal war on Iran. While the self-styled good guys at Anthropic declined to work with the military several months back, the company hasn’t balked at building a predictive surveillance system to monitor activists. Silicon Valley executives have largely dismissed concerns that the data centers fueling their technologies threaten everything from local aquifers to electricity rates, and have seemed relatively unconcerned about the ways that LLM-generated slop is rapidly distorting people’s ability to think and reason independently.
The trouble with trusting AI companies to decide what’s good and what’s bad is that, like other CEOs, Altman and his colleagues have a financial interest in the answers to these moral questions. There are obvious commercial upsides for the executives in continuing to portray their companies (their “labs,” as they like to call them) as upstanding research entities, and mere stewards of a fundamentally mysterious, powerful, and inevitable technology. But they have just as much to gain from the public conversation about LLMs orbiting around the question of whether individual CEOs are good people. It doesn’t really matter one way or the other. They run large and powerful companies that sell products and have no legal obligation to the public good. Those companies appear to be committing crimes, which they could be tried for, using laws that already exist. As far off as real enforcement and regulation might seem, such laws are the only things that will keep the rest of us from being subject to their lousy ethics.
