There is a chart inside the Morning Consult AI Trust Report that almost nobody is talking about, and it may be the most important thing published about AI adoption this year. It shows two lines moving in opposite directions. Trust in AI peaked in October 2025 at a net score near 12 and has been sliding since, landing at 8.5 by April 2026. In that exact same window, self-reported chatbot use climbed from 31% to 49%. The lines cross. Trust goes down. Usage goes up. And the business world, almost uniformly, is drawing the wrong conclusion from it.
Key Facts
- Trust trajectory: Net AI trust peaked at ~12 in October 2025, declined to 8.5 by April 2026
- Usage trajectory: Self-reported chatbot use rose from 31% (May 2024) to 49% (April 2026)
- Distrust drivers: Misinformation concerns (39%), job displacement fears (38%), data privacy (33%)
- Utility trust: Expected AI impact is net positive in healthcare (+14), daily life (+7), education (+5)
- Self-trust finding: Americans trust themselves to use AI responsibly at +16; AI companies at -14
- Source: Should You Actually Care About AI Search?
The divergence is not a contradiction. It is a diagnosis. What it reveals is that there are two entirely separate conversations happening under the label of "AI trust," and most people are treating them as one. The first conversation is about AI as a cultural force: the layoffs, the deepfakes, the data scraping, the concentrated power sitting in the hands of a few companies whose leaders have become polarizing public figures. That conversation has been getting louder since late 2025, and the trust numbers reflect it. The second conversation is about whether AI is useful when someone sits down to do a task. That conversation has a different answer entirely, and the usage numbers reflect that one.
What makes this divergence so strategically significant is not the gap itself. It is that the gap is widening and nobody seems to be updating their behavior in response to it. Marketers are citing the trust decline to justify not investing in AI search visibility. Executives are waving at the 63% distrust figure as evidence that their customers are not using AI to make decisions. Meanwhile, 49% of American adults are using chatbots with some regularity, up 18 points in two years, while trust in the companies behind those tools has gone in the opposite direction. The product is growing. The brand is eroding. And the customer is apparently fine with that arrangement.
This is not a new pattern in technology adoption. People used Facebook for years while distrusting Facebook as a company. They used Google while harboring real concerns about what Google knew about them. The utility of the tool and the reputation of the institution that built it have always been separable judgments, and consumers have always been capable of holding both simultaneously. What is striking about AI is that the separation has happened so fast and so completely. The trust decline is being driven by cultural anxieties, job displacement fears (38%), concerns about misinformation (39%), and data privacy (33%). None of those concerns are about whether the tool helped someone draft a better email or find a vendor faster. They are concerns about what AI is doing to the world, not what it is doing for the individual using it.
The strategic implication is uncomfortable for anyone using trust data to justify inaction. If your customers are making purchasing decisions with the help of AI tools, their institutional distrust of those tools has essentially no bearing on whether you show up in the results. The question is not whether they trust AI. The question is whether they are using it before they call you. At 49% regular usage and rising, the odds are increasingly good that they are. Waiting for trust to recover before investing in visibility is the equivalent of waiting for people to feel good about Google before optimizing for search. The behavior is already there. The sentiment is running behind it, as sentiment almost always does.
Subjective
The divergence thesis is the one I find myself returning to in almost every client conversation, because it reframes the whole debate. The trust question is a red herring. Not because trust does not matter, but because it is measuring something different from what most people think it is measuring. When someone tells a pollster they distrust AI, they are expressing a view about an industry and its cultural footprint. When that same person opens ChatGPT an hour later to research their next software purchase, they are expressing a different kind of judgment entirely: this tool is useful, I trust my own ability to filter what it gives me, and I am going to use it. Both things are true at the same time. The chart with the crossing lines is not a puzzle to solve. It is a map of exactly where the opportunity sits.