Voters Demand Stricter Government Oversight of AI Systems

Voters of both parties want tighter AI regulation, poll finds
An overwhelming majority of likely voters support mandatory formal safety reviews for powerful artificial intelligence systems before they are released to the public, according to a new survey from the AI Policy Institute. The findings indicate a significant bipartisan appetite for stronger government oversight, diverging from the current Trump administration policy that relies on voluntary, opt-in reviews for new advanced models.

Bipartisan Support for Oversight

The survey, conducted June 10 and 11 among 1,007 likely voters, reveals that more than half of respondents across the political spectrum support government-led safety testing. Notably, Republicans showed higher enthusiasm for these measures than Democrats, marking a shift from earlier trends where Republicans were more skeptical of government intervention in the technology sector.
Data from the poll suggests that while voters want regulation, they prefer guardrails over total prohibition. When presented with the choice between banning AI systems or requiring companies to implement safety measures, two-thirds of respondents favored the implementation of guardrails. However, if the only alternative to a total ban is an unregulated AI environment, voters strongly prefer an outright ban.

Demanding Accountability from AI Companies

A significant majority of participants—over 60% of both Republicans and Democrats—believe the federal government, rather than the AI companies themselves, should set clear safety standards and evaluate whether firms are adhering to those rules. Currently, most safety guardrails are designed and implemented by the companies developing the technology.
Public concern regarding the capabilities of future models is also high. More than 80% of respondents, including 84% of Democrats and 83% of Republicans, stated that AI companies should not build systems smarter than humans until they can demonstrate that they can control those systems.

Regulatory Tensions and Data Centers

The survey comes as the federal government navigates complex decisions regarding the release of advanced models. In June, President Trump signed an executive order targeting the cybersecurity of advanced AI, though the resulting vetting mechanism remains voluntary. Meanwhile, the administration has engaged in disputes with companies like Anthropic and OpenAI over the release of their latest models, at times forcing limited releases to trusted partners rather than the general public.
Regulatory debates have also extended to the physical infrastructure supporting AI. The AIPI survey found that 47% of respondents would allow the construction of data centers if they included safety and security requirements, while 38% favored a complete ban. This reflects a broader national trend, as the data center boom faces hundreds of moratoriums and project delays across the country.
As the debate continues, Peter Wildeford, director of policy at the AI Policy Network, noted that the current push for a "no-rules" perspective in some government circles appears to be out of step with the American public. Nearly three-quarters of survey respondents believe that AI will become an increasingly important political issue in the future.