How much do shoppers trust AI shopping assistants?

RTB House reports that consumer trust in AI shopping assistants has already crossed a threshold most retailers haven't planned for. By its account, "39% of U.S. shoppers (and 47% of U.S. millennials) say AI understands their purchase needs better than friends and family," and "across the surveyed geographies, leading AI tools have a markedly higher trust score than established channels such as newspapers, major media outlets, and social media." RTB House says this trust "extends to a meaningful group of consumers who are ready to let AI move from advisor to buyer through autonomous AI purchasing." It names this group "super-trusters": consumers "ready for AI to buy for them, even if returns protection and a seven-day return window are not in place." RTB House is explicit that this segment is a minority — "for most respondents, the removal of these features led to a clear decline in trust" — but frames the existence of super-trusters at all as "the most striking finding," more surprising to RTB House than the majority's safeguard-seeking behavior.

RTB House's own explanation is comparative: shoppers are weighing AI against alternatives — friends, family, newspapers, media outlets, social media — and on its survey's trust metric, AI already wins that comparison for a meaningful share of respondents, concentrated among millennials. The super-truster finding follows from the same logic pushed further: if trust in AI already exceeds trust in human advisors and legacy media for some people, a subset of them will extend that trust to the point of removing transactional safeguards altogether, treating returns protection as unnecessary insurance rather than a prerequisite for comfort.

This sits beside Does trust in AI chatbots drive news-seeking behavior?, which found, across 45 markets, that trust in AI chatbots for news is lower in absolute terms than general trust in news (20% versus 37%) even though trust predicts chatbot adoption more tightly than it predicts social-media use. RTB House's claim is the inverse comparison in a commerce context: AI already scores above newspapers, media, and social media on trust, not below general trust in those channels. The two findings are not strictly contradictory — one measures trust in AI-for-news against trust in news overall, the other measures trust in AI-for-shopping against trust in media channels as information sources — but they illustrate how domain changes the comparison baseline, a caution this library's trust notes should carry forward. The super-truster segment also extends the overtrust pattern in Does co-design participation hide misalignment in preference agents?: where that study found overtrust emerging from a sense of representation despite measurable agent shortcomings, RTB House's super-trusters appear willing to waive concrete financial protections on the basis of general trust alone, a more consequential form of the same dynamic.

The excerpt gives no definition of how "trust score" was measured, no country-by-country breakdown despite claiming the pattern holds "across the surveyed geographies," and no size for the super-truster segment — only that it is smaller than the safeguard-seeking majority. As with the companion note on this source, RTB House sells advertising and retargeting technology to retailers and has a commercial interest in portraying consumer trust in AI commerce as high and rising. The 39%/47% figures are self-reported comparative judgments ("understands needs better than"), not evidence that AI recommendations are in fact more accurate than advice from friends and family, and "trust score" comparisons against newspapers and social media describe stated perception, not a validated measure of trustworthiness. What the excerpt supports, at that strength, is that a meaningful minority of shoppers — disproportionately younger — already rate AI above traditional information sources and human advisors on trust, and that this minority's comfort with removing safeguards is a design signal brands need a tiered response to, not a uniform one.