Headline findings
Six takeaways from our national survey.
People want the legal right to interact with another person, rather than an algorithm, in medicine, in court, in schools, and in benefits offices. They also want disclosure: tell them when they're speaking with an AI, label AI-generated media, and ban the impersonation of people's voice and face.
When AI policy moves to taxing the technology or paying people displaced by it, both proposals still draw majority support, but with thinner margins.
As in earlier national surveys, Republicans and Democrats feel similarly about AI. What separates the AI optimists from the pessimists is how often they use the technology.
We asked whether AI will make work more or less meaningful. Among skilled daily users, about half expect work to become more meaningful. Among less frequent users, the most common expectation is no change.
Whether Americans trust AI a great deal, not at all, or somewhere in between, they support AI regulation at roughly the same level.
Respondents split into thirds by an overall trust-in-AI score; bars show each tertile's average support for AI regulation. (How this is computed.)
About half of U.S. adults expect AI to widen inequality over the next decade; about one in six expect it to narrow. Asked who will gain the most power as AI advances, more than four in ten Americans name large technology companies, and about one in five name AI systems themselves.
