Pew Research Center released a study in June 2026 that should make every commerce team stop and think. About half of US adults now use AI chatbots like ChatGPT, Gemini, or Copilot. That is up from 33% in 2024. Roughly a quarter use them daily. And 42% say they use chatbots specifically to search for information.
That is the adoption curve of a mainstream technology. AI chatbots have crossed the chasm.
But the same study reveals a paradox that defines the next phase of consumer AI. Americans are deeply skeptical of the technology they are rapidly adopting. Majorities say AI is advancing too quickly. Majorities believe AI will put their personal information at risk. More adults predict AI will have a negative rather than positive impact on their lives and on society.
People are using a technology they do not trust. For general information search, that is a concern. For commerce, it is a structural problem.
The Search-to-Purchase Pipeline
When 42% of chatbot users are searching for information through AI, a significant portion of those searches are product-related. “What are the best wireless headphones under $100?” “Which espresso machine is reliable?” “Is this brand legit?” These are commerce queries even when no purchase happens immediately.
The consumer is asking an AI for a recommendation. The AI processes the query, searches available data, and produces a confident, detailed answer. The consumer then acts on that answer.
Here is where the trust paradox becomes dangerous. The consumer does not fully trust the AI. But they act on its recommendation anyway, because the answer is detailed, sounds authoritative, and is easier than doing the research themselves. The AI’s confidence creates a veneer of reliability that the consumer’s skepticism cannot fully resist.
And the AI’s confidence is often misplaced. Not because the model is broken, but because the data the model accesses is corrupted.
The Marketplace Data Problem
When an AI assistant searches for “best wireless headphones under $100,” it pulls data from the open web. The dominant source is Amazon. Amazon’s product listings, star ratings, review counts, and search rankings are the inputs the AI uses to formulate its recommendation.
Those inputs are manipulated at scale.
Fake reviews inflate star ratings. Sponsored listings dominate top search positions. Review counts are boosted by incentivized five-star campaigns. Reference prices are fabricated to create artificial discount signals. Sellers game Amazon’s algorithm to maximize visibility, not to surface the best product.
The AI processes all of this as structured data. It does not know which reviews are fake. It does not know which rankings reflect advertising spend. It synthesizes a confident recommendation from corrupted inputs. The consumer receives a recommendation that sounds well-researched but is actually an amplified version of marketplace manipulation.
This is the core problem. The consumer trust paradox is not just about whether people trust AI. It is about whether the data the AI uses is trustworthy. And right now, for commerce queries, it is not.
Browser AI Makes It Worse
Today, Google started rolling out Gemini in Chrome to users in the UK. British Chrome users can now access Google’s chatbot directly in the browser to compare information across tabs, summarize content, and ask questions about web pages. It has been over a year in development, and it is still not available in the EU. But it is live, and it is coming.
This matters for commerce because the browser is where shopping happens. When a consumer is on an Amazon product page and can ask Gemini “is this product actually good?” without leaving the page, the AI has direct access to the manipulated data on that page. It reads the fake reviews, sees the inflated rating, notes the sponsored ranking, and produces a summary that legitimizes the manipulation.
The consumer asked a trust question. The AI gave a confident answer. The answer was based on untrustworthy data. The consumer is worse off than if they had no AI at all, because the AI’s confidence replaced their own skepticism.
Browser-integrated AI is a power multiplier for whatever data it processes. When the data is good, that is transformative. When the data is manipulated, it is dangerous. It turns a consumer’s healthy doubt into misplaced confidence.
The FTC Recognizes the Problem, Partially
On July 1, 2026, the FTC voted 2-0 to seek public comment on a proposed policy statement addressing AI accuracy. The statement argues that AI companies which distort their systems’ outputs to achieve undisclosed objectives could be deceiving consumers under Section 5 of the FTC Act. The comment period runs through July 31.
The FTC’s framework is focused on ideological manipulation: AI systems that secretly suppress certain viewpoints or promote others. That is a real concern. But the same principle applies to commerce. An AI system that recommends products based on undisclosed commercial incentives is engaging in the same deception the FTC describes.
When an AI recommends a product with 8,000 five-star reviews without disclosing that 3,000 are fabricated, it is deceiving the consumer. The gap between what the consumer expected and what they got is exactly the kind of deception Section 5 prohibits.
The FTC’s accuracy framework needs a commerce layer. Ideological manipulation is one threat. Commercial manipulation through corrupted marketplace data is another, and it affects far more consumers far more often.
What Consumers Actually Need
The Pew data tells us what consumers want. They want AI that works. They want AI that is accurate. They want AI that does not put their information at risk. And they want AI that is not advancing so quickly that it outruns its own reliability.
For commerce specifically, consumers need three things:
Transparent data sources. When an AI recommends a product, the consumer should know where the recommendation came from. Was it based on Amazon rankings? On verified reviews? On an independent trust score? Today, the consumer has no way to tell. The AI presents its recommendation as if it were original analysis, when it is often just a summary of the first page of Amazon results.
Review authenticity before recommendation. The AI should not process reviews without filtering for authenticity. Fake reviews should be excluded before the AI forms its opinion. Today, no major AI assistant does this. They ingest raw marketplace data and process it at face value. Every fake review shapes the recommendation as if it were genuine.
Independent quality scoring. The AI should rank products by actual quality, not by marketplace visibility. A product with 500 genuine, positive reviews and a high quality score should rank above a product with 8,000 reviews (3,000 fake) and a high Amazon ranking. Today, the opposite happens, because the AI uses Amazon’s ranking as a proxy for quality.
How GoBuy Closes the Trust Gap
GoBuy exists to solve the problem the Pew data exposes. Consumers are using AI for commerce queries but cannot trust the results. The root cause is corrupted marketplace data. The solution is an independent trust layer that filters corruption before the AI sees it.
GoBuy’s Smart Score (0-100) is computed from review authenticity, sentiment depth, seller reputation, price-to-quality ratio, and cross-platform consistency. Fake reviews are filtered before the score is calculated. Products are ranked by genuine merit, and only the top 7 per category are returned. The GoBuy Verified badge requires a Smart Score of 80 or higher sustained over 90 days, which means short-term review campaigns cannot game the system.
For consumers using AI assistants, GoBuy’s MCP server at gobuy.ai/api/mcp provides the trust layer. Any MCP-compatible AI agent can call GoBuy’s tools to search products, get trust scores, and compare options with filtered, verified data. The agent gets data the marketplace cannot manipulate. The consumer gets a recommendation worth trusting.
For consumers browsing Amazon directly, the GoBuy Chrome extension injects a trust panel on product pages showing the Smart Score, flagging suspicious reviews, and providing an independent quality assessment alongside Amazon’s manipulated data.
This is how you resolve the consumer AI trust paradox. You do not ask consumers to stop being skeptical. You give their tools data that earns their trust.
The Market Opportunity
The Pew data reveals a market ready for trust-first commerce tools. Half of US adults use AI chatbots. 42% use them for search. And majorities do not trust AI with their personal information or believe it is advancing too fast.
That is not a market that wants more AI capability. It is a market that wants AI they can rely on. The companies that win the next phase of consumer AI will not be the ones with the most powerful models. They will be the ones with the most trustworthy data.
Stop trusting AI that trusts Amazon. Start using tools that verify before they recommend. Visit gobuy.ai to install the Chrome extension, or integrate our MCP server at gobuy.ai/agent-docs to give your AI agents data worth trusting.