Poll Shows Young Conservatives More Willing to Give AI Control Over Policy and Military

In brief

A poll of 18–39-year-old voters found conservatives most open to AI-run policy decisions.
Think tanks and studies show major AI models carry political bias.
Researchers say distrust in institutions may drive the trend.

A new Heartland Institute and Rasmussen Reports survey showed that younger conservatives were more willing than their liberal counterparts to give advanced artificial intelligence sweeping control over major government functions. The findings come despite years of complaints from conservatives that AI systems lean left politically, making the level of support all the more striking.

The poll surveyed 1,496 likely voters aged 18 to 39 between October 31 and November 2. It asked whether respondents supported proposals to let an AI control public policy decisions, determine constitutional rights, or command the world’s largest militaries with the express purpose of “reducing the number of people who die from war.”

Donald Kendal, director of the Glenn C. Haskins Emerging Issues Center at the Heartland Institute, a conservative public policy think tank known for its work on state-level legislation and its criticism of federal regulatory overreach, said the ideological breakdown surprised him.

“If you go into the cross tabs, those who self-identified as Republicans or conservatives were more likely to say yes or support these proposals that are outlined in the survey,” Kendal told Decrypt. “As for why that’s the case, honestly, I’m at a loss.”

The survey relied on self-identified ideology, and those who identified as conservatives showed the highest levels of support in each scenario. The survey reported a margin of sampling error of plus or minus three percentage points.

Kendal said some young conservatives may be reacting to a broader dissatisfaction with existing institutions.

“We’ve got so little trust, faith in our institutions,” he said. “There’s such a terrible approval rating of Congress that it’s so bad that we might as well just blow it all up and start from scratch.”

The Heartland/Rasmussen findings come at a time when trust in government is near record lows. An October 2025 Gallup poll showed that only 15% of Americans approve of the way Congress is handling its job. These low numbers, Kendal said, could explain why some respondents may view AI as a better choice to lead the nation.

“There’s this misconception that AI systems are just these objective sources of truth,” Kendal said.

The results are somewhat confounding. Since ChatGPT’s debut in 2022, researchers and think tanks have documented political bias in major AI systems. A peer-reviewed Public Choice study found that large language models produced left-of-center responses in standardized tests. The Manhattan Institute reported similar results, noting that ChatGPT framed Democratic positions more favorably than Republican ones.

And a 2024 American Enterprise Institute review of GPT-4, Claude, and Gemini found they rated right-leaning think tanks lower in “objectivity,” “morality,” and “quality” than left-leaning institutions.

The Heartland survey also found that more than a third of young voters supported giving an AI system control of the world’s largest armies. Kendal said some respondents may have taken the question’s framing at face value.

“If you’re taking that in good faith, fewer casualties of war is a fairly sympathetic dream,” he noted. Kendal said voters’ thinking could be that “if putting AI in charge of this could reduce the amount of wars, or at least the amount of carnage associated with these wars, then, let’s just do it like it’s certainly not working out the current way.”

Kendal said the support levels underscored a misunderstanding about how AI systems work.

“One of the things I try to drive home is dispelling this illusion that artificial intelligence is unbiased. It is very clearly biased, and some of that is passive,” Kendal said, noting that increasingly society is relying more heavily on AI.

He added: “We do so at our own peril and with a blindfold on, because these things aren’t obvious.”