While most Americans are now comfortable with letting AI into their daily lives, many don’t think it's useful enough to pay for. A new study from HR tech firm Howdy of more than 1,000 Americans found that just 26% are currently paying for an AI platform, while 54% said they’d drop an AI tool if they had to pay for it. At first blush, the findings are positive for the big AI firms: Adoption is very high, with 96% of respondents reporting that they’ve tried AI, and 86% saying they use it routinely. Americans are also starting to trust AI more than other outlets: 34% of respondents said they believe AI results over what they see on social media, 20% trust AI more than what the government tells them, and 14% hold AI results in higher regard than what journalists report. But Americans aren't yet paying for AI because while they generally regard AI tools as useful, they don’t see them as transformative. Howdy’s researchers found that while 90% of Americans think AI is a useful search tool, only about half (51%) rely on it for planning—and only 18% say that AI is useful in making them money. With AI increasingly becoming a fixture in day-to-day life, it’s no surprise that skepticism and anxiety about the technology has grown with its adoption. After ChatGPT launched in 2022, 63% of Americans favored a ban on artificial general intelligence, according to nonprofit research group Sentience Institute. As a result of these fears, 68% of Americans now believe that investing in humans is more important than further investment in AI, and 70% think that building data centers is “not worth the environmental impact.” The survey found that some AI platforms have bigger image problems than others. Claude, Gemini, and ChatGPT are viewed more favorably, with trust rates of 12%, 21%, and 42%, respectively.  Not so for Grok, the least trusted of the lot—or, as the study’s authors phrased it, the one in the “doghouse.”