University of Toronto Study: Earned Media Dominates AI Search

Summary

This post examines a University of Toronto study that reveals how AI search engines are redefining authority across industries. It compares platforms like ChatGPT, Perplexity, and Gemini against Google and shows that AI engines consistently prioritize earned media while diminishing or removing brand-owned content and social platforms. Automotive loses consumer forums, electronics sidelines influencer-driven content, software shifts from vendor pages to independent validation, and local services face fragmented results. These findings matter because they confirm that AI search is rewriting the rules of visibility. The path to presence now runs through expert reviews, trusted publishers, and machine-readable data rather than brand blogs or social chatter. This post highlights why marketers and communicators must treat earned media as central to their strategy if they want to remain visible in AI-driven search.

The University of Toronto recently published a study that exposes how AI search engines handle citations across industries. This is not my data. It is original academic research conducted with the help of ktau.ai, a platform used for large-scale search analysis. The research tested platforms like ChatGPT, Perplexity, and Gemini against Google to understand which sources appear in answers. The findings show dramatic shifts that marketers, PR leaders, and analysts cannot ignore.

AI search does not treat all industries equally. In fact, the study found that automotive, consumer electronics, software, and local services each produce distinct patterns of source citations. Google tends to mix brand-owned, social, and earned media. AI engines, by contrast, lean heavily toward earned content such as expert reviews and publisher sites. The result is a new distribution of authority that alters how visibility is achieved.

This change matters because the engines are shaping perception at every stage of the customer journey. A product recommendation pulled from a trusted review site carries more weight than a brand’s own website or a community forum thread. In industries where AI excludes social platforms altogether, the voice of the consumer disappears from discovery pathways. That silence changes the competitive game.

As you read through this breakdown, keep one point in mind. These are not projections or guesses. They are insights directly from empirical tests conducted by University of Toronto researchers. They show us the emerging rules of AI visibility, and they make clear that brands must adapt their strategies to survive in this new search environment.

Automotive: From Community Talk to Publisher Authority

The automotive industry revealed some of the sharpest contrasts in the University of Toronto’s research. Google balanced its results across brand-owned sites, social platforms, and earned media. In Canada, for example, 36.6 percent of citations were brand, 22.8 percent were social, and 40.6 percent came from earned outlets. The pattern in the United States was similar. Google pulled heavily from brand sites and social platforms like Reddit, while still featuring expert reviews.

AI engines told a different story. They pushed brand content down and removed social sources almost entirely. In Canada, AI search delivered 69.1 percent earned and 30.9 percent brand, with zero social citations. In the United States, the trend was even stronger. AI answers leaned 81.9 percent earned and 18.1 percent brand. Social platforms were excluded completely.

That shift shows where authority is being reassigned. In automotive, consumer conversations in forums and communities are no longer a factor in visibility through AI. Authority sits almost entirely with expert-driven review sites and publisher outlets. For car makers, this signals a major change. Traditional local SEO and brand-controlled messaging will not carry the same weight. The fight for visibility is happening in editorial coverage, trusted reviews, and expert analysis. If your brand cannot win presence in those spaces, the AI engines will push you off the shortlist.

Consumer Electronics: Reviews Rule the Rankings

Consumer electronics showed an even sharper pattern. According to the University of Toronto data, Google delivered a mixed ecosystem. In Canada, 54.1 percent of citations came from earned outlets, 23.1 percent from social, and 22.8 percent from brand-owned sites. In the United States, Google leaned more toward brand content at 32.9 percent and social at 15.4 percent, while earned media still carried nearly half of the weight.

AI search engines shifted the balance decisively. In Canada, 77.6 percent of citations were earned, 22.1 percent brand, and only 0.3 percent social. In the United States, the tilt was even more extreme. AI answers leaned 92.1 percent earned, almost eliminating social sources, with the remainder coming from brand-owned content.

The data makes one truth very clear. AI engines are deprioritizing social platforms like Reddit and YouTube, even though those communities have long shaped consumer decisions in electronics. Instead, authority is being handed almost entirely to professional reviews and publisher-driven outlets. For electronics brands, this means the battleground has shifted. The forums, unboxings, and influencer chatter that once dominated visibility now sit outside the AI’s field of view. If your brand is not present in expert reviews or trusted editorial sites, you will struggle to appear in the AI’s shortlist of recommendations.

Software Products: Independent Validation Over Vendor Claims

Software produced another striking contrast. According to the University of Toronto’s findings, Google leaned heavily on brand-owned sources. In Canada, 53.8 percent of citations came from brand sites, 14.4 percent from social, and only 31.8 percent from earned media. In the United States, the numbers shifted slightly but told the same story. Google produced 43.7 percent brand, 10.9 percent social, and 45.4 percent earned.

AI engines reversed the pattern. In Canada, 74.2 percent of citations were earned, 25.8 percent brand, and no social content at all. In the United States, the split was 72.7 percent earned, 26.7 percent brand, and negligible social presence. In other words, AI engines pushed vendor-owned sites to the margins and amplified independent assessments from third-party reviewers and publishers.

Software Products: Independent Validation Over Vendor Claims

This is a crucial shift. In software, brand-controlled pages have traditionally carried significant weight in Google rankings. AI engines no longer see them as authoritative by default. Authority now belongs to expert reviewers, analyst outlets, and independent publishers. For software companies, this means credibility must be earned outside their own domain. Vendor blogs and landing pages will not be enough to win a spot in AI-generated answers. The new game is about independent validation, trusted citations, and authoritative third-party coverage.

Local Services: A Fragmented Landscape

Local services introduced an entirely different challenge. The University of Toronto study showed that overlap between Google and AI engines was very low in this category. In home cleaning, roofing, tax preparation, and dentistry, overlap rarely broke above 20 percent. In auto repair and IT support, overlap was nearly nonexistent.

This means AI engines and Google are surfacing entirely different providers in local searches. Google still favors directories, local SEO signals, and branded service pages. AI engines, on the other hand, are inconsistent. They often pull from scattered sources, and in many cases, their answers exclude providers that dominate Google’s local rankings.

For marketers, this signals a new frontier. Traditional local SEO—optimizing Google Business Profiles, collecting reviews, and maintaining directory listings—will not guarantee visibility in AI search. Local businesses must now think differently. Structured data, machine-readable service details, and coverage in authoritative review outlets become essential. The path to discovery is fragmented, and winning visibility means playing by AI’s rules, not just Google’s.

Cross-Category Insights: A Systematic Shift in Authority

When looking across industries, the University of Toronto research revealed three consistent patterns. First, AI engines overwhelmingly favored earned media in every vertical and every region. Expert reviews, publishers, and institutional sites carried far more weight than brand-owned or social sources. Second, Google maintained a more balanced ecosystem, still lifting brand-owned content and social platforms in many categories, especially consumer-facing ones like electronics and automotive. Third, AI engines almost entirely excluded social platforms, silencing consumer conversations that once influenced discovery pathways.

For marketers, the lesson is direct. AI search narrows the definition of authority, making third-party validation the deciding factor for visibility. Brands that invest in winning coverage across expert outlets, reviews, and credible publishers will rise in AI rankings. Those that rely on brand-owned content or community chatter will fade. The playing field has shifted, and the winners will be those who treat earned media as the cornerstone of their strategy.

Conclusion: The New Rules of Visibility

The University of Toronto research makes one point undeniable. AI search engines are rewriting how authority is defined and how visibility is achieved. Each industry tells the same story in different ways. Automotive lost social influence, electronics pushed community voices aside, software sidelined vendor-owned content, and local services fractured into inconsistent results. Across all of them, earned media emerged as the dominant source of credibility.

For marketers and communicators, this is more than a shift in rankings. It is a reset of strategy. The fight for visibility now takes place in expert reviews, credible publisher sites, and structured data that machines can easily read. The days of relying on brand blogs and directory listings are fading fast. What wins in AI search is trust, validation, and authority from third-party voices.

The takeaway is clear. You need to build strategies that treat earned media as a primary channel, not an afterthought. You need to think of your website as an API for machines, not just a brochure for people. And you need to accept that AI engines will decide visibility in ways that Google never did. The brands that understand this shift will capture presence in AI-generated shortlists. The ones that ignore it will vanish from the answers consumers trust most.