The Reuters/Ipsos poll finds 73% of Americans believe AI companies haven’t done enough to prevent serious societal harm. Public negativity toward AI has increased: 39% now view it as harmful and only 11% view it positively. Most respondents favor government-led safety standards, 55% support slowing AI development, and awareness of calls for restraint by leading firms is high. The online survey sampled 1,277 U.S. adults (±3 percentage points).
Poll: 73% of Americans Say AI Firms Aren’t Doing Enough to Prevent Harm
Most Americans are worried AI companies aren't doing enough to prevent serious harm hero image
A Reuters/Ipsos poll released Tuesday finds broad public concern about the safety of artificial intelligence: 73% of Americans say AI companies have not done enough to prevent the technology from causing serious harm to society.
The online survey, conducted over four days and completed Sunday, collected responses from 1,277 U.S. adults and carries a sampling error of plus or minus 3 percentage points. The poll shows rising negativity toward AI: 39% of respondents now view the technology as harming society, up from 36% a month earlier. Just 11% described AI as a positive development, with the remainder unsure or declining to answer.
Public Wants Government-Led Safety Standards
Respondents expressed widespread skepticism that AI companies can adequately police themselves. A strong majority said the federal government should take the lead on safety: when asked to choose priorities, 73% prioritized safe and responsible AI development over maintaining U.S. dominance in the AI race (23%).
The poll also found 55% of respondents believe slowing the pace of AI development would be beneficial, while 13% said slowing would be harmful. Awareness of public discussion around restraint was high: 69% of Americans reported seeing media coverage in which leading AI firms called for a slower development pace and stronger regulation.
Notable voices: Earlier this month a researcher at Anthropic left the company, saying they feared AI development could lead to outcomes devastating for humanity. The scientist, Hubinger, estimated the odds of such a catastrophic event occurring within the next decade at greater than 10%.
Legislation and Political Context
Progress toward comprehensive federal AI legislation remains limited. Reuters notes that voices across the political spectrum — including U.S. Senator Bernie Sanders and Steve Bannon — have urged federal action. At the same time, Donald Trump has publicly championed rapid AI development. The Republican-controlled House was in recess through November at the time of the poll, a factor that may slow legislative momentum.
The Reuters/Ipsos findings align with other Ipsos data showing two-thirds of Americans believe AI is developing too quickly, and nearly three in four say the government should act to prevent AI-related job losses. As of June 2026, just 38% of Americans described the AI news coverage they see as favorable, down from 44% a year earlier.
Methodology: The online poll sampled 1,277 adults across the United States and reports a margin of error of ±3 percentage points.
Help us improve.
