Kelly has been especially active in AI policymaking, with a demonstrated interest in harnessing its possible benefits while also addressing risks
Sen. Mark Kelly, D-Ariz., talks with reporters in the Senate subway on Tuesday, December 9, 2025. (Tom Williams/CQ-Roll Call, Inc via Getty Images)
Earlier this month, city officials in Chandler, Arizona unanimously rejected a $2.5 billion artificial intelligence data center supported by ex-Sen. Kyrsten Sinema. Then Tucson officials shot down an Amazon $AMZN data center three months before. For Democratic Sen. Mark Kelly, it reinforced a belief that tech companies must switch up their playbook if they want to prevent a wider revolt against data center construction.
“I think the community does get a say and they showed that in Chandler, that they get to decide,” Kelly told Quartz on Thursday. “And I think what happened in Chandler and in Tucson shows that companies need to work with these communities upfront.”
Only 44% of Americans would welcome a data center in their community, according to a September survey from the Heatmap Pro intelligence platform. Local resistance was quickly evident once plans for a 422,000-square-foot data center reached Chandler’s doorstep, despite a projection of $2.5 billion in new economic activity. Residents were more unnerved about fresh noise pollution, draining the local water supply, and ultimately paying more in monthly utility bills.
The Chandler planning office reported 200 comments against the data center. Only eight were in favor. Kelly, though, believes that further developing AI can be accomplished while ensuring locals aren’t on the hook and paying more for electricity.
“If we see that become a trend across the country, it's going to be hard to win against China in this race to develop intelligent models where we can reap some significant benefits in scientific discovery and cures for diseases and material science and chemistry,” Kelly said.
Compared to his other Democratic colleagues in the Senate, Kelly has been more active in AI policymaking with a demonstrated interest in harnessing its possible benefits while simultaneously addressing risks. He’s one of 15 members of the Senate’s bipartisan AI caucus and also up for re-election next year.
In September, he rolled out a 24-page AI roadmap that included a novel idea: Tax frontier AI companies if they kill jobs, and use the new revenue to set up retraining programs for displaced workers and upgrade aging infrastructure like the power grid.
It's named the “AI Horizon Fund,” the product of discussions with stakeholders among universities and industry stakeholders like OpenAI, the maker of ChatGPT.
“I want to see AI be successful. I think this is a real opportunity for the American people if we get it right,” he said. “But we got to get it right. If we don't get it right, we're going to have a lot of unemployed people with no opportunity.”
Among Democrats, fears of mass job displacement from AI are starting to spread. Kelly cited a 2023 estimate from the McKinsey Global Institute that projects 12 million people could be displaced from their jobs by 2030 as a result of AI. He isn’t alone in the upper chamber.
Virginia Sen. Mark Warner has observed the unemployment rate among new college graduates in the labor force has climbed to 9%, rivaling the levels reached during the Great Recession. He warns that figure could soar to 25% in the near-future. Warner — normally soft-spoken and reluctant to use profane language — recently said on social media that possibility “should scare the sh*t out of us.”
In Kelly’s case, he keeps tabs on AI’s impact on the job market and likes to use ChatGPT occasionally. The Arizona Democrat recalled a conversation with Duolingo CEO Louis Von Ahn, who relayed to him his belief that the firm’s software engineers can only become more productive from AI with no threat of layoffs.
If it doesn’t pay off, Kelly won’t tip his hand on whether a bailout is necessary, a scenario floated by OpenAI in November. “I generally think government bailing out big billion dollar corporations because they made bad decisions, we always want to try to avoid that,” he said.
Kelly does see opportunity for bipartisan cooperation on AI, despite Trump's attempts to unilaterally enact a ban on state AI rules. The Arizona Democrat said he has not spoken with Sinema, his former colleague and now a lobbyist, in recent months.
He did list Republican Sen. Josh Hawley’s bill designed to crack down on chatbot companions for minors as one he was considering co-sponsoring. He also said he was working alongside Sen. John Curtis of Utah on a new bill to regulate social media algorithms. "These are real issues people are concerned about."
Join 500,000+ readers who start their day with Quartz.
By subscribing, you agree to our Terms of Service and Privacy Policy.
