The public relations industry stands at a critical crossroads in 2026, grappling with its identity in an era dominated by Large Language Models (LLMs) and Search Generative Experiences (SGE). As brands increasingly demand to know how they appear in the outputs of artificial intelligence, a specific statistic has taken hold within the communications sector: the claim that earned media drives 84% of all AI citations. However, a deeper analysis of the underlying data suggests that this figure is not only misleading but potentially damaging to the long-term credibility of the PR profession. If the industry is to provide strategic counsel on AI visibility, its own foundational data must withstand the same rigorous scrutiny it applies to client messaging.
The Genesis of the 84% Figure
The statistic currently circulating through PR agencies and boardroom presentations originates from a comprehensive study conducted by Muck Rack titled "What Is AI Reading?" Released in May 2026, the report analyzed a staggering 25 million citations across 17 different industries, making it one of the largest datasets of its kind. The study aimed to identify the primary sources of information used by chatbots to generate responses for users.
While the dataset is robust, the interpretation of the "84%" headline requires significant nuance. The study focused on three primary chatbots: OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Gemini. It found that a vast majority of the citations provided by these platforms could be categorized under a broad umbrella of "earned" or "third-party" sources. However, the controversy lies in the categorization. The 84% figure bundles together a diverse array of sources—including academic research, government publications, Wikipedia entries, Reddit threads, and social media posts—under the single banner of "earned media."
For a PR professional, "earned media" typically refers to editorial coverage secured through media relations efforts. When the data is disaggregated to focus specifically on journalism and news media, the figure drops precipitously. Analysis suggests that traditional journalism accounts for approximately 25% of AI citations—a significant portion, but far from the dominant majority suggested by the headline.
The Disconnect Between Data and PR Strategy
The danger of the 84% myth lies in the strategic advice it fosters. When PR teams present this figure to clients, the implicit suggestion is that an increase in traditional media relations will lead to a direct and proportional increase in AI visibility. This creates an expectation that "winning the GEO (Generative Engine Optimization) race" is simply a matter of scaling up press release distribution and journalist pitching.
In reality, the most cited sources across the major LLMs remain non-journalistic entities. Wikipedia, PubMed, and Reddit consistently outperform even the largest global news organizations in terms of citation frequency. These are platforms that PR teams rarely influence through standard media relations. By overselling the impact of journalism on AI outputs, agencies risk misallocating client budgets toward tactics that address only one-quarter of the AI visibility puzzle.
Furthermore, the Muck Rack study omitted some of the most frequently used AI search surfaces, such as Google’s AI Overviews and Perplexity. These platforms utilize different ranking signals than standard chatbots, often prioritizing real-time data and transactional information that differs from the training data used by Claude or older versions of ChatGPT.
The Search Engine Fallacy: SEO vs. GEO
The confusion surrounding AI data is further compounded by recent consumer research. A July 2026 report from YouGov, "Searching for Answers: How AI is changing online discovery in 2026," surveyed 2,003 UK adults regarding their digital habits. The report found that 85% of respondents still use "search engines," while only 31% claimed to use "AI assistants."
At first glance, these numbers suggest that traditional Search Engine Optimization (SEO) remains far more important than Generative Engine Optimization (GEO). However, this conclusion ignores the fundamental shift in how search engines operate. On May 19, 2026, during the annual I/O conference, Google announced that "AI Mode" would become the default experience for all search queries. This effectively merged the concepts of search and AI assistance.
When a user performs a Google search today, they are interacting with an AI-driven interface that synthesizes information from across the web. Therefore, the 85% of users who claim to be using a "search engine" are, in fact, using AI. The distinction between a "search engine" and an "AI assistant" has become a semantic one, yet many in the PR and marketing industries continue to treat them as separate silos. This lack of technical understanding leads to flawed reporting and a failure to recognize that GEO is not a future trend, but the current reality of SEO.
Chronology of the AI Information Shift (2024–2026)
To understand how the PR industry reached this point of statistical confusion, one must look at the rapid evolution of the AI landscape over the past 24 months:
- Late 2024: Major LLMs begin integrating real-time web browsing capabilities, moving away from static training sets. PR teams start noticing brand mentions in ChatGPT but lack tools to measure them.
- Early 2025: The term "GEO" (Generative Engine Optimization) enters the marketing lexicon. Agencies begin experimenting with "citation mining" to see which articles are being picked up by AI.
- May 2026: Google I/O marks a turning point. Google declares that "Google Search is AI Search," making AI Overviews the primary way users consume information.
- May–June 2026: Muck Rack and YouGov release their respective reports. The PR industry quickly adopts the "84%" figure as a defensive metric to justify media relations budgets in an AI-first world.
- August 2026: Industry analysts begin to point out the discrepancies in the data, noting that the broad definition of "earned media" used in studies does not align with the actual work performed by PR agencies.
Analysis of Implications for the PR Industry
The reliance on inflated or misinterpreted data poses a systemic risk to the PR industry. If agencies continue to promise that media relations will dominate AI citations, they will eventually face a "day of reckoning" when client dashboards fail to show the promised results.
The actual impact of journalism on AI is significant, but it is part of a complex ecosystem. LLMs prioritize:
- Authority: High-domain authority sites like Wikipedia and government portals.
- Consensus: Information that is repeated across multiple reputable sources (including, but not limited to, news).
- Community Validation: Platforms like Reddit where "human" perspectives are synthesized.
- Structured Data: Academic journals and technical documentation.
For a PR strategy to be effective in 2026, it must move beyond the "hit" and focus on the "footprint." This means ensuring that a brand’s narrative is consistent across news media, third-party review sites, Wikipedia, and even niche community forums.
Reactions from the Field
Data scientists within the marketing sector have called for a more "objective view" of LLM inputs. David Clare, Director and Head of Fusion at Fire on the Hill, has been a vocal critic of the industry’s current reliance on the 84% figure. His firm’s internal tracking, which monitors over 3 million data points across platforms like ChatGPT, Claude, and Google AI Overviews, suggests that while earned media is a vital ingredient, it is being oversold.
"If we want brands to trust our guidance on AI visibility, the data behind it needs to meet the standard we set for everyone else," Clare noted in a recent industry analysis. He argues that by ignoring the other 75% of factors that influence AI responses, PR professionals are doing a disservice to their clients and the industry’s reputation for accuracy.
Toward a New Standard of AI Measurement
The PR industry must pivot from defensive metrics to transparent, evidence-based reporting. This involves several key shifts in how success is measured:
- Refining Definitions: Agencies must distinguish between "Journalistic Earned Media" and "Broad Third-Party Sources." This transparency allows clients to see exactly where their visibility is coming from.
- Integrating SEO and PR: The total convergence of search and AI means that PR teams can no longer work in isolation from SEO departments. The technical health of a website is now just as important for AI visibility as a feature in a major newspaper.
- Focusing on Information Integrity: As AI engines become more adept at spotting misinformation, the PR industry’s role as a guardian of factual accuracy becomes more valuable. Spreading misleading statistics about AI to sell PR services is fundamentally at odds with this role.
In conclusion, the "84% myth" serves as a cautionary tale for an industry in transition. While it is tempting to embrace statistics that validate traditional methods, the reality of the AI era is far more complex. The PR industry’s future depends on its ability to navigate this complexity with honesty, technical proficiency, and a commitment to data integrity. Only by acknowledging the true nature of AI citations can the industry provide the sophisticated counsel that brands require in this new digital frontier.
