Is AI going to kill us all? We asked nuclear engineer (and Democrat) Chris Jones

Chris Jones, the nuclear engineer-turned-Democratic nominee running against French Hill in Arkansas’s 2nd Congressional District, has also become a prominent voice on artificial intelligence. Jones gave a talk a few months back on both the opportunities and risks artificial intelligence could pose for the Arkansas Times. He also recently published a 13-part Substack on AI as well as issued a three-part roadmap for how to prevent the AI apocalypse. 

His timing is prescient: the heads of AI companies — as well as disaffected former employees — have sounded the alarm in recent weeks, warning that AI could, in fact, kill off humanity. 

In the edited video interview below, we had the chance to ask Jones under what scenarios AI could wipe out civilization and whether he thinks the risks to humanity are overblown. Jones said that threats to humanity are not new.  “At every moment in the history of humanity, there has been a concern for annihilation because humans have a capacity to destroy civilizations,” he said.

That said, Jones is taking the warnings very seriously, comparing the threats from AI to those of nuclear technology.  

“I’m a trained nuclear engineer, right? And so the likelihood that a nuclear reactor will melt down is, generally speaking — because of the guardrails — not that high. But if a nuclear reactor does melt down, the impact and danger is massive. Look at Chernobyl. It’s long-lasting. And so AI is similar,” Jones said. 

AI programs are increasingly capable of hacking into systems — and making independent decisions. Would an AI system press a button to launch missiles or harm civilians? Once AI systems can design and train themselves — a development some experts believe is just around the corner, Jones noted — that possibility could become real. 

But the candidate isn’t an AI doomer who thinks humans are destined to obliterate or enslave ourselves at the hands of our chatbot overlords. He sees a lot of potential benefits of artificial intelligence, even in Arkansas, particularly for sectors like education, agriculture, rural healthcare and small businesses.

Understaffed hospitals, in particular, Jones said, could leverage AI to treat more patients. “You can have AI diagnostics that strengthen rural hospitals,” said Jones. “So, now, instead of having to go on doctor’s visits, AI can take specific patient data and come up with a diagnosis that a doctor can affirm — because you don’t want to take humans out of the loop — but it’ll do it in half the time and be able to say, ‘Here are the specific medicines that will address your specific concerns.’” 

But the benefits of AI come with an obvious downside: job loss. Jones thinks estimates that AI could eliminate a third or more of existing jobs could be accurate, with massive implications for Arkansas’ workforce. There are things politicians can do to help, he said.

In some countries, he noted, companies can’t lay off workers and replace them with AI. In the U.S., some politicians are calling for the spoils of AI to be shared across the population, through proposals for an AI public wealth fund, for example, given that chatbots and large language models are built on our data. 

Although President Trump has resisted calls for international cooperation on AI, Jones said he would push Congress to work across borders, encouraging countries to create collective guardrails for the technology — naming the Nuclear Non-Proliferation Treaty as an example — and collaborate on research. While kumbaya cooperation with other countries may seem like a fantasy as long as Trump is in office, Jones said he’s taking the long view. 

“He won’t be in office forever, but AI will be around long after he leaves office,” Jones said. “I think we can’t give up just because right now there’s someone who is against it, because the threat is too massive.”