In the United States, there is a noticeable cooling toward artificial intelligence, and this is no longer just isolated sentiment but a sustained trend confirmed by fresh data from my latest analysis. According to the results of a large-scale survey conducted among the adult population, 39% of respondents are now convinced that AI does more harm than good to society. For comparison: a year ago, this figure was 31%, and in 2023 it reached 40%. Meanwhile, 52% of respondents assess the technology's effect as balanced, and only 9% see it as an unequivocal benefit.
Paradoxically, the rise in skepticism is accompanied by increased awareness. In 2026, 70% of American adults stated that they are at least superficially familiar with AI, whereas in 2024 that figure was 64%. This suggests that the deeper society delves into the topic, the more wary it becomes — a classic technology maturity cycle that we observed with the internet and social media.
Young people lead in disappointment
The most dramatic shift occurred in the 18–29 age group. Among them, 47% now consider AI rather harmful (a year ago — 36%). Trust in business regarding the responsible application of the technology has collapsed: only 20% of young people are willing to trust companies (down from 30%), and 41% completely reject corporate initiatives (up from 29%). This generation, which grew up with technology, now sees its costs better than anyone.
Anxiety about the labor market has also intensified to record levels. 79% of respondents expect job losses due to AI in the coming decade (in 2025 — 73%). This is felt especially acutely by middle-aged people: in the 45–59 age group, the figure jumped from 75% to 84%. Overall, only 27% of Americans trust corporations in matters of ethical AI implementation, which is 4 percentage points below last year's level.
The study was conducted online among 3,270 adult residents of the United States between May 4 and May 11, 2026. Notably, public sentiment is already extending beyond surveys: in July, about 200 activists in San Francisco held a protest between the offices of OpenAI, Anthropic, and Google DeepMind, demanding a halt to the training of more powerful models.
My comment: These figures are a signal for the industry. We are entering a phase where technological optimism gives way to pragmatic regulation. Ignoring these sentiments risks not only reputational losses but also accelerated legislative restrictions that could slow down innovation. The balance between progress and public trust will become the main challenge for market leaders in the coming years.