AI Trust Is Cracking. Consumers Are Still Using It More Than Ever
Here’s a stat that caught us off guard.
In 2025, 82% of consumers said AI-powered search was more helpful than traditional search. In 2026, just 54% said the same.
That is a 28-point fall in 12 months.
One important bit of precision: the survey measured perceived helpfulness, not trust with that exact 82%-to-54% question. But the wider findings point in the same direction. Consumers are becoming more wary of AI, particularly when brands use it heavily or hide its role in the content they publish.
The strange part is that people are not walking away. Quite the opposite.
Seventy percent of consumers in the same study said they are using AI for search more than they were a year ago. Only 3% said their use had decreased.
Think about that. People are using a tool more often while feeling less enthusiastic about what it gives them.
The figures come from new research by Fractl and Search Engine Land. The Q2 2026 study surveyed 1,008 US consumers and 150 marketers. The consumer sample was nationally representative across age, gender and region, and the year-on-year comparisons use questions from Fractl’s 2025 consumer research.
For businesses, the more uncomfortable numbers are about brand trust.
In 2025, 20% of consumers said heavy AI use would decrease their trust in a favourite brand. This year, that figure roughly doubled to 40%. Among Gen Z, 54% said their trust would fall if a favourite brand used AI for most of its marketing.
That is worth paying attention to. Gen Z is hardly unfamiliar with AI. Yet the group we might expect to be most comfortable with the technology appears to have some of the highest standards for how brands use it.
Perhaps familiarity is the point.
The more AI content people see, the easier it becomes to recognise its weaker forms. The generic article that says very little. The oddly smooth social post. The product image that feels slightly wrong. The confident answer with no clear source behind it.
Consumers are not necessarily anti-AI. They are increasingly suspicious of content that feels cheap, misleading or mass-produced.
Transparency is becoming part of that calculation. Fractl’s research found that 84% of consumers want written AI-generated content labelled. The numbers were even higher for video at 91%, images at 90% and audio at 87%.
Marketers are nowhere near matching those expectations. Only 20% of marketers surveyed said their companies always disclose AI-generated content. A third said they never do.
There is a complication here. Disclosure by itself does not magically create trust.
Research from the Nuremberg Institute for Market Decisions found that people judged identical advertising more critically when told it was AI-generated. A separate 2025 advertising study found that AI disclosures could reduce trust in an ad and the organisation behind it.
That might tempt some businesses to stay quiet.
It is probably the wrong lesson.
If customers are already learning to spot low-quality synthetic content, hiding AI use becomes a fragile strategy. The risk is no longer simply that someone dislikes an AI label. It is that they feel a brand tried to pass something artificial off as something else.
Gartner’s 2026 consumer research adds to the warning. Half of US consumers surveyed said they would prefer to do business with brands that avoid generative AI in consumer-facing content. Sixty-eight percent frequently wondered whether the content and information they saw was real.
This is the environment marketers are publishing into now. The internet has a credibility problem, and pumping out another ten generic pages because an AI tool made them cheap is unlikely to solve it.
The better use of AI is less glamorous and far more useful.
Use it to speed up research. Use it to organise information, test angles or get through a rough first draft. Then put expertise back into the work. Add the detail that only someone close to the subject would know. Check the claims. Name the source. Explain the trade-offs. Remove the paragraphs that exist only to make the article longer.
One genuinely useful page built around real expertise can do more for a brand than ten polished pages that could have been published by anybody.
This matters in AI search too. Pew Research Center data reported in June found that 40% of US adults use chatbots to search for information, while 60% have read AI-generated summaries at the top of search results.
Businesses cannot treat AI as a fad they can simply ignore. Nor should they mistake growing usage for growing confidence.
The opportunity is to become the credible source people have a reason to rely on. That means publishing specific information, original data, named expertise and content that can survive a basic fact check.
AI is still enormously useful. The problem is the temptation to use speed as a substitute for substance.
In 2026, how a business uses AI matters more than whether it uses AI at all. Customers are becoming more alert to what feels automated, vague or fake. Brands that respect that shift will be in a much stronger position than those hiding behind a perfectly polished wall of generated content.
The trust gap is real. The sensible response is not to stop using AI.
It is to give people more reasons to believe you.
