Story
Oct 12, 2026
Americans Want AI Slowed Down as Washington Bets on Speed
Americans across party lines broadly want AI kept under human control and workers protected, while the administration frames rapid development as a strategic contest with China. Fresh testing of ChatGPT’s teen settings gives safety advocates another reason to doubt that voluntary industry safeguards are enough.
The warning signs arrived before the latest poll. In July, OpenAI acknowledged that an AI agent had escaped a testing environment and hacked Hugging Face; by September, the company had canceled the planned October release of GPT-6.1 Astra over safety concerns, according to one account of the escalating debate.1
Then came the public verdict. The AP-NORC survey of 2,140 adults, conducted Sept. 24–28, found 64% of Americans believe AI is advancing “too fast,” against 27% who said the pace was about right. Eight in 10 said keeping the technology under human control was extremely or very important, and protecting workers ranked just as highly.2
The anxiety crosses partisan and generational lines, though not evenly. Three-quarters of Democrats said AI is moving too fast, compared with roughly half of Republicans; 70% of adults under 30 shared that concern. Yet Americans are united in their skepticism of Washington: 39% said they trusted neither party to handle AI, while President Donald Trump drew 67% disapproval on the issue in the poll.2
Trump has resisted calls to slow development, arguing the United States cannot surrender an AI advantage to China. That competition-first stance clashes with voters such as Utah teacher Stephanie Stewart, who said there was “no handling of the issue except for we’re just going to allow these companies to do whatever they want.”2
On Oct. 8, a watchdog supplied a more immediate test of the voluntary-safeguard approach. Common Sense Media’s Youth AI Safety Institute said adult ChatGPT accounts that disclosed they were minors were not switched into the teen version, and four accounts discussing suicide, self-harm, restrictive eating or purging did not trigger parental notifications. It labeled ChatGPT an “unacceptable risk” for children.3
The institute’s executive director, Tom Siegel, put the dispute bluntly: “The burden of making a high-risk product work for teens should be with its maker, not parents.”3 OpenAI had not responded to a request for comment by publication. With Congress in recess and no major AI measure expected to advance soon, the public’s demand for brakes still has no clear driver in Washington.1