When Rick Perry thinks about AI, he thinks about nuclear weapons

It’s pretty easy to conjure up dystopian visions for how the global pursuit of artificial intelligence will play out. After all, governments, militaries, and startups frequently push ideas seemingly ripped from the pages of a Philip K. Dick novel, or worse. It’s easy to think of the pursuit of AI, for instance, in the oft-used frame of an arms race. It’s so easy that US Energy Secretary Rick Perry, ostensibly at a press conference to tout the positive impact of his department’s AI research portfolio, pretty quickly started talking about nuclear weapons.

At the Department of Energy AI event in Chicago, Perry took a question on the risks of the United States not leading the world in artificial intelligence. The Oct. 2 event was hosted by Argonne National Laboratory, which will soon take possession of the United States’ first exascale supercomputer—a really powerful computer with all sorts of artificial intelligence applications. After touching on the competition between the United States and China in supercomputing, Perry, standing alongside Energy Department Under Secretary for Science Paul Dabbar and others, reached for metaphors from the annals of US nuclear history to talk about just how important artificial intelligence is.

“Paul and I happened to be together at Hanford yesterday in front of the B Reactor that was the first reactor in the world that produced plutonium, the material that the Trinity project and then the bomb at Nagasaki that helped end World War II were produced there.” said Perry, before ending on a more upbeat note. “Where I’m going with this is that it’s the innovators, the innovators in America, and I don’t want to get too far afield here, but it’s one of the reasons it’s so important for America to invest in STEM programs, to excite young people into science and technology.”

But surely AI and the competition between the United States and China to develop it isn’t the same as the Cold War-era race to develop nuclear weapons. Asked to clarify, Perry seemed to simultaneously double down on and walk back his previous words. “What I’m saying is that artificial intelligence is a lot more than just the weapons program—it’s a part of it. I would say that, and Paul you would probably agree with me here, the ancestor to AI that we see today goes back to the old weapons program, the ability to compute all of the science that was required to build the B Reactor,” Perry said. “It’s so vast in how it’s going to affect all of our lives.”

Noteworthy views on artificial intelligence aside, Perry’s been in the news for other reasons lately. Congressional Democrats are questioning the energy secretary’s ties to Ukraine in the context of the impeachment inquiry against President Donald Trump. On Friday, Trump reportedly blamed Perry for pushing him to make the fateful call to the president of Ukraine that jump-started the impeachment inquiry. Numerous outlets have reported Perry will be resigning by year’s end.

Perry’s certainly not alone in viewing artificial intelligence in Cold War or martial terms. Earlier in the day, US Rep. Will Hurd talked about the need for US dominance in artificial intelligence. China, he said, wasn’t using cutting-edge technology like facial recognition to “make it easier to buy groceries.” But there’s a danger to framing competition in developing artificial intelligence as an arms race; such framing, many experts (including the University of Cambridge’s Heather Roff) say,  creates the risk that governments will adopt military policies that make AI-based conflict more likely.

Besides, China probably has made AI that helps you buy groceries.