Elon Musk talked with President-elect Donald Trump before the launch of a SpaceX Starship rocket on Tuesday. Brandon Bell/Associated Press One of the Bostonians sounding the alarm about the dangers of unbridled artificial intelligence is Russ Wilcox, a venture capitalist and former tech entrepreneur. He helped shepherd “electronic ink” technology from the labs of MIT to the screens of e-book devices such as the Amazon Kindle. Wilcox’s firm, Pillar VC, has invested in AI startups. But he’s also concerned about AI’s potential to render huge numbers of white-collar workers irrelevant and eventually evolve into a “new species,” with its own goals and an insatiable need for power. Elon Musk has been sounding similar warnings about AI for a long time. He’s said it could prove more dangerous than nuclear weapons. He was an early funder of OpenAI, which was originally set up as a nonprofit with a mission to serve “the best interests of humanity.” Advertisement And now Musk may be able to influence President Trump on the subject. Will Musk be a voice of caution about AI? Or will he be more concerned with getting the government to use AI developed by his own company, X.ai — possibly as part of his new effort to reduce waste in the government? AI investors like Wilcox are watching closely, so I asked for his perspective. This interview has been condensed and edited. Trump wants deregulation everywhere. How do you square that with what Elon Musk — and you — believe are the risks of totally unchecked AI development? Musk has been warning about the dangers of AI, and he supported a bill in California that was going to provide regulation. So to the extent Trump has got Musk in his ear, the policy of the US government may be a flip of a coin every single day and every single tweet. But my guess is we’ll be in a bitter struggle or race with China to see who can define the breakthrough AI, and we’ll be trying to keep them from having chips, and we’ll be trying to invest in our own efforts. I’m not saying it’s the wrong policy for the US. If we’re going to be in a race, we should win the race. But a race is not a safe thing for the species. Advertisement A better result would be diplomacy, where we all slow down. But I just don’t think that’s what’s going to happen. Do you think Trump will undo Biden’s executive order from 2023 that put some guardrails on AI or a more recent one that stopped American venture capitalists from investing in Chinese AI companies for national security reasons? There’s no chance Trump will cancel the executive order to stop American VCs from backing Chinese AI. But I do guess he will cancel Biden’s AI executive order, because that adds a safety review step and he is generally trying to kill rules that limit business. The person who could best convince him to take AI safety to heart would be Musk. As you know, Musk was an early backer of OpenAI, and Musk does worry about AI safety. There have been various UN task forces on AI — the risks, the biases, equality of opportunity, the typical things that the UN thinks about. But is there a scenario where you could see the UN creating some guardrails and regulations that everybody agrees to, to get us out of that race condition that you describe? I think that would be a good result. I’m not optimistic. The US likely will come out of the Paris Climate Accords. That’s another one of these thorny problems where we all have to cooperate. I just think that the new administration’s attitude is “Let there be no rules on the US, and leave us alone.” Musk is running an initiative that has just been named, and nobody quite knows where it sits or what authority it has — the Department of Government Efficiency. Do you think Musk will suggest that AI replace a lot of government workers? Advertisement The cost for a human to do office work is $50 an hour, $30 an hour. Depending on how much energy you’re going to use and how smart the office work is, 10 cents is probably the high end of what we’re going to spend to have an AI do the same office work that a bureaucrat does. I do think AI will replace a lot of white-collar office work. Probably not inside the next four years, because I just think it’s going to take people a couple years to build the tools. The mass economic transformation is probably five or 10 years out. Russ Wilcox is operating partner at Pillar VC, which is based in Boston. Courtesy of Russ Wilcox For now, though, how do you see Musk navigating what could be fewer government regulations around AI while he’s a government vendor with SpaceX and he now has an AI company, X.Ai, that competes with OpenAI? In the absence of regulations, Musk will feel just fine exercising his own judgment about AI safety and charging ahead. That guy definitely thinks he is smart enough to play with the future of the human race. Plus, he might score some fat government contracts for his own AI company that will give him a shot to spring into the lead. But will his AI be safe for society? What we saw in the recent election is that Musk manipulated the algorithm on X [the former Twitter], according to a study, to favor his political allies, and he personally posted all kinds of misinformation and tropes. I find that behavior menacing. I don’t think we want Musk to be the one creating AI without some bipartisan oversight and regulations in place. Advertisement