Survey shows GenAI at 41% in US, while agentic AI lags

GenAI adoption in the US has accelerated to 41% of adults as of June 2026, with many users turning to AI for search-like tasks rather than creative work.

New survey findings show awareness is also spreading quickly: just 12% of US adults say they have not heard of GenAI, down sharply from 58% in March 2023.

The same June 2026 results indicate a large “middle” audience remains. About 30.5% said they have heard of GenAI but do not understand it, down from 35.7% a year earlier.

Search and summaries dominate day-to-day GenAI use

Among current users, the most common activity resembles a replacement or supplement for traditional web search. “Searching on the Internet” ranked as the leading use case, cited by 53% of GenAI users in June.

Close behind, 48% said they use GenAI for research-related support, including generating summaries and surfacing relevant information. The prominence of these two categories suggests the technology is being adopted first for low-friction, information-seeking tasks.

By contrast, several widely discussed consumer use cases are not expanding in the same way. Education-related use was reported by 22.1% of GenAI users in June, slightly below 22.5% in March 2026 and down from a 24.3% high in March 2025.

Content creation also slipped. The share using GenAI to draft articles, blog posts, or social media posts was 17.1% in June, down 0.8 percentage points quarter over quarter and 1.6 points below the three-year average cited in the survey figures.

Privacy, accuracy, and transparency concerns remain high

Even as usage expands, consumer anxiety about data practices is intensifying. In June, 61% of adults said they were extremely or very concerned about AI using their data, up 0.5 points quarter over quarter and roughly 4 points versus December 2025.

Concerns about incorrect outputs are also climbing. The share worried about AI giving wrong information rose to 40%, up from 25% when the question was first introduced, while 40% additionally said AI should be overseen by humans.

Broader social and governance worries show up alongside these product-level issues. In June 2026, 36% said they were concerned AI could contribute to job losses, and 33% called for more disclosure about what data AI systems rely on.

Regulatory sentiment is edging upward as well. The portion saying GenAI should be regulated by the government reached 28%, rising 1 point from the prior quarter and 4 points year over year.

Underlying many of these views is a trust gap. In June, 29% of respondents said they do not believe AI has their best interests in mind, up 0.5 points quarter over quarter and 1.25 points compared with September 2025.

Agentic AI struggles to win consumer permission

While mainstream GenAI use has become common, “agentic AI” remains far from broad acceptance. The June 2026 survey found 68% of US adults have never heard of agentic AI, while only 8% said they already use it.

Enthusiasm is limited: 10% said they were excited to experience agentic AI. When asked whether the concept is a good idea, just 18% agreed, while 42% disagreed and 41% were unsure, with “no” increasing versus the previous quarter.

Behavioral questions reinforce the same pattern. When offered a set of real-world tasks for AI agents, 60% chose “none of the above,” indicating reluctance to delegate decisions or transactions.

The most accepted tasks still registered in low double digits: 12% would use an agent to find and book a restaurant, 11% to purchase groceries, and 10% each for smart-home management and paying bills.

The near-term implication is a two-speed consumer market: GenAI continues to spread for information retrieval and summarization, while autonomous “do it for me” agents face a higher bar tied to privacy, accuracy, and user control. What comes next will likely depend on clearer data safeguards, stronger transparency, and product designs that keep humans in the loop for consequential actions.