A University of Pennsylvania survey of 1,330 Americans reveals deep AI pessimism and a bipartisan majority demanding stronger federal regulation of artificial intelligence.
Only 17% of Americans believe artificial intelligence will have a positive impact on the country over the next decade. That number, from a fresh nationally representative survey of 1,330 U.S. adults conducted February through March 2026, is striking on its own. Paired with the 42% who expect AI's effects to be negative, it adds up to a public that has watched the technology's rapid rollout and landed somewhere between doubt and alarm.
The poll, conducted by the Annenberg Public Policy Center at the University of Pennsylvania, captures sentiment across party lines. Sixty-five percent of respondents said the federal government has done "too little" to regulate AI, a view held by majorities of Democrats, independents, and Republicans alike. The margin of error is ±3.49 percentage points. "The demand for regulation is not a partisan issue," said Shawn Patterson Jr., a research analyst at APPC. "Majorities across the political spectrum say the government has done too little."
What makes the skepticism harder to dismiss is who holds it. Seventy-eight percent of respondents said they have heard at least a moderate amount about artificial intelligence, and 67% reported using AI tools at least a few times in the past month. This is not a public reacting in ignorance to something abstract. These are regular users who have formed an opinion based on experience, and that opinion skews negative.
The product deluge driving that awareness has been relentless. Forbes noted earlier this year that ChatGPT, now three years old, had accumulated 800 million monthly active users and $20 billion in annual recurring revenue, one of the fastest consumer adoption curves in technology history. OpenAI is pressing that advantage by nearly doubling its headcount to 8,000 by year's end, concentrating new hires in product, engineering, and sales, CNBC reported in March, citing the Financial Times.
Other players are embedding AI just as aggressively into hardware and social platforms. Google unveiled the Googlebook this week, a new laptop line with Gemini replacing the traditional cursor with an AI-activated Magic Pointer, according to MacRumors. A hands-on test by PCWorld found the feature promising but unfinished. The product pipeline is moving faster than anyone's ability to evaluate it.
That context matters for reading the survey results. Americans who say they are using AI regularly and still expect negative effects are not reacting to hypothetical scenarios. The concern is concrete enough to map onto a real political demand: two-thirds want more government action, and they want it regardless of party.
The regulatory backdrop
Congress has repeatedly stalled on comprehensive AI legislation. The previous administration's executive order on AI was partially reversed in early 2025, and state-level efforts have produced inconsistent frameworks that differ widely across jurisdictions. The result is a policy vacuum the industry has largely filled on its own terms.
The Annenberg data follows a familiar pattern in technology regulation history. Support for social media oversight built slowly after years of visible harms, eventually producing congressional hearings and a handful of targeted laws. With AI, the timeline may be compressed. The technology is more directly integrated into daily tasks than algorithmic content ranking ever was, and the public appears to be forming hard opinions faster. That the regulatory demand is already bipartisan and majority-level, before any single dominant domestic AI incident, suggests the political window for legislation may be wider than the current congressional record implies.
Whether Washington moves before the next product cycle does is an open question. But 65% is a durable number.
---
Frequently Asked Questions
What share of Americans expect AI to harm the country?
According to the Annenberg Public Policy Center's February-March 2026 survey, 42% of Americans expect artificial intelligence to have a negative impact on the U.S. over the next decade, against just 17% who expect a positive effect.
Is concern about AI regulation bipartisan?
Yes. The survey found that majorities of Democrats, independents, and Republicans all said the government has done "too little" to regulate AI, making it one of the few areas of broad political agreement.
How many Americans are actively using AI tools?
Sixty-seven percent of survey respondents reported using AI at least a few times in the past month, and 78% said they had heard at least a moderate amount about the technology.
Who conducted the survey and how large was the sample?
The Annenberg Public Policy Center's Institutions of Democracy division surveyed 1,330 U.S. adult citizens from February 17 to March 20, 2026, with a margin of error of ±3.49 percentage points.