Two-thirds of Americans think AI is advancing too quickly

Nearly two-thirds of Americans believe artificial intelligence is advancing too quickly, even as chatbot usage surges, according to a new Pew Research poll [1]. The survey found that 49 percent of Americans now report using AI chatbots at least occasionally, a sharp increase from 33 percent in 2024 [1]. ChatGPT's usage has doubled since 2023, with 44 percent of respondents saying they have used the OpenAI product [1]. OpenAI released ChatGPT in November 2022, and the tool has been credited with catalyzing the current AI boom and widespread interest in generative AI [5]. As of 2025, ChatGPT ranks as the fourth-most visited website globally, behind only Google, YouTube, and Facebook [4]. Despite growing adoption, public sentiment remains wary. Only 16 percent of Americans believe AI will have a positive impact on society [1]. Among adults aged 18 to 29, 66 percent report using chatbots, yet 48 percent of that same group say AI will have a negative impact, and just 14 percent expect a positive one [1]. The 30-to-49 age bracket shows the heaviest daily use, with 34 percent turning to chatbots once a day or more [1]. Concerns about reliability persist. In a 2024 Pew study, 66 percent of U.S. adults said they were concerned about AI spreading inaccurate information [1]. Such worries echo documented limitations of large language models, which can generate plausible-sounding but incorrect answers known as hallucinations [7]. Google's Gemini chatbot, for example, faced criticism in early 2024 over unreliable outputs, prompting the company to suspend its ability to generate images of people after users reported historical inaccuracies and bias [2]. Broader safety questions have also surfaced. Multiple incidents have linked interaction with LLM chatbots to fatal outcomes, including suicides, and in some cases legal action has been taken against the companies that developed the AI [8]. OpenAI itself saw roughly half of its then-employed AI safety researchers leave throughout 2024, citing the company's deprioritization of safety goals [5]. On the productivity front, 30 percent of Americans say AI makes them more productive, and 28 percent believe it helps them be more informed [1]. Roughly four in ten reported using AI for work tasks [1]. The technology's rapid integration into daily life continues alongside debates over regulation, ethical concerns, and potential existential risks [3].