It could be the end of the world as we know it. Or maybe it’s all just a marketing stunt.
Somehow, these have become the two poles of the debate about the future of artificial intelligence and its potential perils. While the argument has been going for a while, it reached a fever pitch when a researcher at leading AI developer Anthropic resigned and issued a public warning that recklessly designed AI could literally wipe out humanity (for example by helping bad actors create a virus or other bioweapon, or even by going rogue and attacking infrastructure). Days later, Anthropic CEO Dario Amodei published a lengthy essay essentially calling for an intentional slowdown of AI development, as well as government regulation. Sam Altman of rival OpenAI and Elon Musk publicly agreed.
But what was really notable about this latest round of AI concern is how surprisingly vociferous the skeptics of the doomsday narrative have become. Most strikingly (and theatrically), Nvidia CEO Jensen Huang accepted a call from Donald Trump while on stage at a tech conference, putting the president on speaker to wave away AI fears: “It’s a hoax,” Trump said, amplifying his critique on social media that too much regulation would ultimately be good for China. “The robots are not going to be taking over the world.” Huang nodded along. In a subsequent appearance on CNBC, Huang added that new regulations the AI leaders seem to be seeking are “completely unnecessary.”
This may seem like a curious flex on Huang’s part. While Nvidia, a maker of chips coveted by AI firms, is not a consumer-facing brand, it’s become familiar to many by way of its soaring share price. Yet Huang seems to be shrugging off any reputational fallout from its no-worries stance—even as public opinion of AI (shaped by data center backlashes, job-loss predictions, and other factors) sinks. Political figures from Bernie Sanders to Steve Bannon are calling for curbs on the technology. There have been attention-getting incidents of AI behaving in troubling ways, and many corporate leaders seem genuinely concerned. It seems like a weird moment to go hard on sneering away from AI regulation.
But it’s not just Huang. Tech investors and self-styled seers like David Sacks and Marc Andreessen have hewed to a familiar “no regulations” framework argument. And most recently, Meta’s Mark Zuckerberg waded into the discourse by declaring that leading A.I. labs could do more to prioritize safety more internally, rather than pushing the tech forward at all costs. It was a take largely interpreted as a dig at Anthropic, but clearly framing the issue as one to be resolved by free enterprise, not regulation.
Broadly, the skeptics’ case is that the leading AI players want to shape the regulation they’re ostensibly seeking, with the resulting business upside of locking in their own advantages and kneecapping up-and-coming or future competition (a form of what’s sometimes called “regulatory capture”). AI leaders like Amodei and Altman are accused of playing up safety concerns because they are planning to release safety-focused product updates. “What better way to create demand than to create a problem?” Nvidia’s Huang said recently. “Who doesn’t want their market to be hysterical about their product and line up around the corner for it?”
Good question. And it’s fair to say there’s some de facto branding tied up in the AI labs’ calls for regulation. Anthropic in particular has cultivated an image as a company that’s distinctly careful and thoughtful—and that image certainly involves implying that its technology is both bleeding-edge and monumentally powerful. (And some critics, such as Trump, seem to find CEO Amodei annoyingly smug and high-handed about safety issues.)
That said, it’s hard not to see many of those opposing a go-slow, regulated approach as engaging in a bit of marketing, too. Nvidia’s Huang, for example, is selling the most desired shovels in the AI gold rush, and thus benefits from the largest possible pool of miners. His upside is that demand for Nvidia chips has given his firm a market lined up around the corner, so to speak. Why would he want a slowdown?
So sure, there’s some self-interest in the so-called doomsayers’ rhetoric. But then again, doom marketing isn’t that great a strategy, so maybe they mean what they say just the same. And it’s also fair to wonder if there’s just as much self-interest driving the case for “trust me!” optimism. Denying risk is its own form of risk. And nobody wants to find out just how bad an AI misadventure could get if the optimists are wrong.