A new Pew Research poll has found that nearly two-thirds of Americans believe artificial intelligence is developing too fast, even as adoption of AI tools continues to climb at a striking rate. The Verge reported on the survey, which shows that 49 percent of Americans now use chatbots at least occasionally, up sharply from 33 percent in 2024, while 63 percent of respondents expressed the view that the technology is advancing too quickly.
The tension those numbers reveal is not new, but rarely has it been captured so cleanly in a single dataset. Americans are, in meaningful numbers, integrating AI chatbots into their daily routines — and simultaneously telling pollsters they are uncomfortable with the speed at which the underlying technology is moving. That is not necessarily a contradiction. People adopt tools because those tools are useful, or because professional and social pressure makes non-adoption costly, without necessarily endorsing the pace or the direction of the broader industry that produced them. What the poll describes, in other words, is a public that is participating in an AI-driven shift while harboring genuine reservations about it.
The context matters here. The period between 2024 and now has been one of extraordinary acceleration in the consumer AI space. Chatbot products have moved from novelty to embedded feature across productivity software, search engines, customer service platforms, and smartphones. Companies including Google, Microsoft, Apple, Meta, and OpenAI have competed aggressively to put AI interfaces in front of as many users as possible, often defaulting these tools on rather than asking users to seek them out. That structural push almost certainly accounts for a portion of the usage growth the Pew survey captures. Someone who uses an AI-assisted feature inside a product they already relied on may describe themselves as a chatbot user without having made any deliberate choice to become one.
That makes the adoption figure somewhat harder to interpret than it appears at first. Growth from 33 percent to 49 percent in roughly a year is dramatic by any polling standard, but it is happening inside a deployment environment that has been explicitly engineered to maximize that number. The 63 percent who say the technology is advancing too quickly may, in part, be responding to exactly that feeling of being moved along faster than they chose to move.
For the technology industry, the likely reading of this poll will depend on who is doing the reading. Companies pushing rapid deployment will find comfort in the adoption numbers and argue that use is outrunning anxiety, treating the gap between the two figures as evidence that concerns are ultimately overcome by experience. That is a reasonable interpretation in the short term, but it carries risks. Public ambivalence at this scale, particularly on a technology that is simultaneously affecting employment, media, education, and civic life, has historically created the conditions under which regulatory pressure builds. The fact that a majority of users are also skeptics means the industry cannot easily separate its customer base from its critics.
For policymakers, the survey provides political cover in either direction. Legislators who want to move on AI oversight can point to the 63 percent figure as evidence of constituent concern. Those who prefer a lighter hand can point to the rapid adoption curve and argue that interference would disrupt something the public is clearly choosing to use. The ambiguity is genuine, and it suggests that AI regulation in the United States will continue to be contested and incremental rather than sweeping, at least in the near term.
The Pew findings also raise a question the summary touches on but does not fully resolve, concerning ChatGPT specifically. OpenAI's flagship product has been the defining consumer AI brand since the end of 2022, and its name recognition shapes how much of the public understands the entire chatbot category. Whether ChatGPT's share of usage is growing, shrinking, or holding steady relative to competitors matters for understanding where the industry's center of gravity sits. Without those specifics, the broader usage figures describe a category rather than a winner.
What to watch in the months ahead is whether the gap between adoption and concern narrows or widens. If it narrows, the most likely explanation is normalization — people becoming more comfortable as the technology becomes more familiar. If it widens, the more plausible reading is that the pace of change is genuinely outstripping people's ability to assess it, which historically tends to produce a harder political and regulatory correction. The speed at which AI capabilities are being rolled into everyday products means the public will not have much time to deliberate. That, in a sense, is precisely what two-thirds of them say they are worried about.
