I disagree with the premise. The cited survey says:

    Thinking now about AI (A-I), that is artificial intelligence …
    Q28 In general, would you say… (ROTATE) +
      The benefits outweigh the risks of AI ............. 34
      The risks outweigh the benefits of AI ............. 57
      A little of both (VOL)............................. 2
      Not sure .......................................... 7
    + Results shown reflect responses among registered voter

That's all well and good, but the question is very broad.

I think when pushed, the reason people hate AI is that a lot of "AI" solutions have been rolled out before the technology was mature. Even now, for end-user facing things most applications are pretty bad, even as the models have gotten more powerful.

You're increasingly pushed to use AI by every online service, and every interaction with retail pushes AI in your face, and importantly, ALL OF THEM ARE JUST TERRIBLE. Amazon has their shopping assistant which is, astonishingly, WORSE than their regular search, which is a hell of a bar.

But people who use ChatGPT or Claude to answer questions on their own are thrilled with the technology. They'll occasionally laugh at the R's in strawberry, or try to torture the AI (this might be more of a teenager thing) but it is remarkable how well-liked it is, as opposed to Google's search-integrated AI for example.