Voters Want AI Regulated. Washington Has a Voluntary Pledge and an Empty Chamber.

The numbers arrived this week, and they describe a political vacuum. A Reuters/Ipsos survey of 4,506 U.S. adults, including 3,526 registered voters, found that most Americans believe neither the White House nor Congress is taking the risks of artificial intelligence seriously. Fifty-seven percent of registered voters said the Trump administration had not taken AI risks seriously enough or at all. Fifty-four percent said the same of Congress. The survey was conducted online over six days and ended Monday, October 5; results were published October 7 with a margin of error of two percentage points.

The headline figure is the jobs number. Eighty-four percent of voters said AI is a threat to American workers. For comparison, 60 percent said the same of illegal immigration, 74 percent of competition from foreign workers, and 73 percent of large corporations. AI now tops a list of worker anxieties that politicians have spent decades arguing about. Even fears about the technology slipping entirely out of control are durable: 62 percent said AI could get out of control and risk the future of humankind, essentially unchanged from the same poll in August 2025.

The regulation consensus that Washington cannot convert

Voter demand for rules cuts across party lines. Eighty percent of Democrats and 61 percent of Republicans said they support stricter government regulation of AI, a bipartisan majority in an election year that has very few of them. The leaders calling for federal action stretch across the spectrum too, from Senator Bernie Sanders, the independent who caucuses with Democrats, to Steve Bannon, the hard-right ally of President Donald Trump. Calls that broad usually produce legislation. Here they have produced almost nothing.

Congress has failed to agree on legislation setting guardrails for the technology, and lawmakers are expected to be away from Washington until after the November 3 midterm elections. The administration's signature action to date is a handshake: last week Trump said executives from six leading AI developers, Nvidia, SpaceX, OpenAI, Anthropic, Meta and Alphabet's Google, had agreed to voluntary safety principles, including working with independent auditors to assess whether AI systems behave as their designers intended. The agreement includes no stated consequences if a company chooses not to comply.

84 percent of registered voters call AI a threat to American workers, a higher share than for illegal immigration, foreign worker competition, or large corporations.

The data center front line

If there is one place where abstract anxiety becomes concrete, it is the data center. The survey found 56 percent of voters back limits on building the facilities that power AI, including 69 percent of Democrats, 55 percent of independents and 43 percent of Republicans. Data centers have become a genuine campaign issue: candidates from both parties are urging a pause in the construction boom, which has landed in suburbs and rural towns as noisy, power-hungry neighbors that push up utility bills. Loudoun County, Virginia, already one of the densest data center markets in the world, is now a case study in how quickly infrastructure politics turns local.

Advocacy groups have noticed. On October 5, the conservative AI safety group Humans First drove a bus bearing anti-data-center messages past OpenAI's headquarters in San Francisco during a nationwide tour. When activists from opposite ends of the political spectrum start agreeing on the villain, policymakers usually follow eventually. The keyword is eventually.

A red line on military AI

The poll also found growing unease about AI in warfare. Fifty-seven percent of voters think the U.S. government should not use AI to determine the target of a military strike, up from 49 percent in the August 2025 poll. That eight-point move in fourteen months suggests that reports of AI agents running amok and hacking into corporate networks, alongside warnings from prominent researchers about existential risk, are shifting public tolerance faster than policy is moving. Voters are drawing a distinction between tools that assist people and systems that make life-and-death decisions, and they are drawing it more sharply than any current federal rule does.

Across the Atlantic, a different answer

The contrast with Europe is worth spelling out, because it shows what a regulatory answer actually looks like. On October 9, EU tech chief Henna Virkkunen told Reuters that the bloc's AI Act is already equipped to handle rogue AI systems, saying the law covers the whole life cycle of capable models. The European Commission sent requests for information in late August to more than 30 AI companies about their safety, transparency and copyright obligations, and Virkkunen said she is now assessing the responses. Those requests can lead to investigations carrying fines of up to 7 percent of a company's global annual turnover.

Europe's approach is not a model anyone can copy wholesale. It is expensive, contested, and companies argue it is already outdated. But it is a law, with enforcement machinery, deadlines and penalties. As of October 11, the American answer is a voluntary pledge with no stated consequences, a Congress on recess, and a poll showing that voters know the difference.

The midterms will test the vacuum

AI has become a significant campaign issue ahead of the November 3 midterms, when control of Congress is up for grabs. A White House official, responding to the poll, pointed to a new task force and said the administration's priority is ensuring there is a booming economy so that workers have many opportunities and experience limited volatility from any short-term disruptions. That is an economic answer to a safety question, and the poll numbers suggest voters noticed.

Daniel Kokotajlo, a former OpenAI staffer who left in 2024 over concerns the company was not behaving responsibly, framed the public's hierarchy of worries plainly: abstract existential concern is not top of mind for most Americans, while the fear that AI costs people their jobs or threatens health and safety is. Voters are not asking Washington to solve the distant future. They are asking it to regulate the present. So far the present has given them a pledge, a poll, and an empty chamber.