The premise embedded in this question is partially wrong. Medical and legal categories do not appear less frequently in AI Overviews overall – they appear at rates significantly higher than most industries. What they face is a higher qualification threshold that most sites in these categories cannot clear. Understanding the distinction changes the strategy entirely.
How Google’s Liability Sensitivity Affects AI Overview Source Selection in YMYL Categories
YMYL – Your Money Your Life – is Google’s framework for content where inaccurate information can cause real-world harm. The September 2025 Quality Rater Guidelines update expanded the definition to include elections, institutions, trust in society, and civic dimensions. Four risk categories now fall under YMYL: health and safety, financial security, government and civics and society, and other sensitive topics.
The counterintuitive data: Legal queries trigger AI Overviews at 77.67% – the highest trigger rate of any YMYL category, per SE Ranking’s analysis of 1,200-plus keywords. Health queries trigger them at 65.33%. Finance at 41.67%. Politics and elections at 16.67%. Previsible’s December 2025 analysis found legal category AI adoption is 11.9 times higher than non-YMYL categories, finance 2.9 times higher, and health 2.9 times higher.
The liability sensitivity does not suppress YMYL queries. It raises the E-E-A-T gating threshold for the sources cited in those AI Overviews. Google actively generates medical and legal AI Overviews – it just draws almost exclusively from institutional or credentialed sources, leaving the long tail of non-credentialed YMYL content uncited.
The disclaimer behavior confirms this: 83% of health-related AI Overviews include the “For medical advice or diagnosis, consult a professional” warning. 63.2% of finance AI Overviews include equivalent professional consultation language. The disclaimers indicate the AI Overview is present and active – they accompany the content, they don’t suppress it.
The Specific Content Patterns That Trigger AI Overview Suppression in Medical Topics
Hard suppression in health applies narrowly. Mental health crisis queries are fully suppressed and replaced by the “Help is available” feature directing users to the 988 Suicide and Crisis Lifeline. Specific medication dosage queries show near-zero AI Overview rates. Local provider queries – “dermatologist near me,” “cardiologist in [city]” – dropped to 0% AI Overview coverage by December 2025, routing exclusively to Maps and the local pack.
Outside these suppressed formats, health query coverage is exceptionally high. BrightEdge’s December 2025 healthcare analysis found treatment and procedure queries at 100% AI Overview presence, pain queries at 98%, symptom and condition queries at 93%, and medical coding queries at 90%. Clinical content is effectively fully saturated with AI Overviews. Volume does not determine this: very low-volume health queries under 1,000 monthly searches show 100% AI Overview rate; high-volume queries above 100,000 monthly searches show 89%. Google uses AI for long-tail medical queries regardless of traffic scale.
The dominant cited sources for health AI Overviews: Mayo Clinic at 107 links, YouTube at 104 links, WebMD at 91 links per SE Ranking’s analysis. Institutional authority outweighs general web authority across the board. AI hallucination risk in health is the mechanism: AI produces unsupported medical claims 50% of the time; Stanford research found severe errors in 22% of medical Q&A responses. The E-E-A-T gating is Google’s response to these error rates.
The specific content pattern that triggers exclusion from medical AI Overview citations is not topic proximity – it is credential absence. Content lacking clear author credentials was removed from featured snippets and AI Overview consideration simultaneously. The absence of verifiable author expertise, institutional affiliation, or clinical evidence from peer-reviewed research is the filter.
Why Legal Information Sources Face Higher Citation Thresholds
Legal content has the highest YMYL AI Overview trigger rate in the entire category at 77.67%, but the citation threshold is specifically calibrated to the harm type within legal content. Legal educational queries – “what is habeas corpus,” “how does discovery work in civil litigation” – trigger AI Overviews at high rates. Specific legal advice queries – “should I sue,” “is this non-compete clause enforceable in California” – trigger suppression or single-source citation with a disclaimer.
The hallucination rate is the driver: AI hallucinates court holdings 75% of the time. This severity of error risk requires higher human verification standards. Legal sources cited in AI Overviews must have attorney-authored or reviewed content, verifiable bar registration information, and jurisdiction-specific accuracy. A legal content piece that is factually accurate on general principles but does not identify the authoring attorney, does not specify jurisdiction, and does not carry review dates fails the threshold regardless of its substantive quality.
YMYL expansion in September 2025 broadened this: legislation, legal code, and government official behavior now fall explicitly within scope. Legal publishers covering regulatory content, compliance requirements, and government-adjacent information now face the same credential and verification requirements as those covering individual legal rights and procedures.
The distinction between educational and advisory legal content maps cleanly to suppression behavior. Legal definition content and procedural explanation content earns citations. Specific advice content triggers suppression or single-source citation from official government sources. Most small legal publishers produce a mix of both content types on the same site, which creates a citation pattern where some pages get cited consistently and others never appear – without the publisher understanding why the pattern exists.
The Credential and Verification Signals That Help YMYL Sites Break Through
Seven non-negotiable content elements for AI Overview citation in YMYL categories, per UpGrowth’s analysis: entity authority via E-E-A-T; evidence attribution; clinical or legal source documentation; regulatory compliance signals; transparent limitation statements; clear topic boundaries; and clinical or legal accuracy verification through explicit review processes.
Author verification signals that AI systems recognize: medical professionals with Google Scholar publications; attorneys with bar registrations verifiable on state bar websites; financial advisors with FINRA records; academic affiliations with verifiable institutional pages. These create digital footprints in the knowledge graph that AI systems can cross-reference. A credential that exists only on the publisher’s About page without third-party verification carries minimal weight.
Wikipedia presence is a concrete, underused lever. Building Wikipedia pages for leading physicians on the team dramatically increased AI citation rates in documented case studies. Wikidata entries for the organization are also required – entity recognition by AI systems depends on third-party knowledge base presence, and Wikipedia and Wikidata are among the most-weighted knowledge base sources. An organization that exists in Wikidata with linked authoritative personnel is structurally different from an organization that does not, even if the content quality is identical.
The healthcare case study that illustrates the threshold: a regional network switched from non-medical writers to board-certified physicians authoring content, implemented a peer review process, and added last-reviewed dates to all clinical pages. The result was a progression from Google quality penalties to featured snippet appearances and AI Overview citations. The content topic did not change. The credential and verification architecture changed entirely.
Perplexity’s medical standard is the most stringent benchmark: it returns 21 or more sources per medical answer, prioritizes peer-reviewed literature from PubMed and clinical trial databases, filters unverified and unattributed sources, and boosts academic medical centers and hospitals. Sites targeting Perplexity citations for medical content need PubMed-cited authors or academic center affiliation. Google AI Overviews are less stringent but trending in the same direction.
Realistic AI Overview Expectations for Medical and Legal Content Producers
The traffic paradox in health is documented. Health sites are seeing 40 to 70% traffic drops on informational and symptom pages – pages that are simultaneously being cited in AI Overviews. The AI Overview is consuming the query that previously generated the visit. Appearing in AI Overviews doesn’t restore this lost traffic – it is structurally caused by AI Overview presence. For health publishers, the strategic choice is: invest in AI Overview citation for brand visibility and authority signals, or accept organic traffic cannibalization without the brand presence benefit.
Structural protection zones remain untouched. Local provider queries, physician profiles, service line pages, and location and facility pages receive 0% AI Overview coverage. Traditional SEO on these pages delivers full ROI because there is no AI layer to compete with. These are the patient acquisition pages – the pages closest to actual healthcare decisions – and they are completely outside AI Overview territory.
10.4% of AI Overview citations in YMYL categories are themselves AI-generated content, per Originality.AI’s analysis of 29,000 YMYL queries. This creates a long-term model collapse risk as AI-generated YMYL content cites other AI-generated content. For credible YMYL publishers, human-authored content remains the citation-quality differentiator precisely because the AI-generated content share is measurable and increasing.
Realistic ceiling for credentialed YMYL sites: legal informational content with attorney authorship and bar verification can achieve citation rates comparable to non-YMYL informational content. Health content with physician authorship, institutional affiliation, and peer-reviewed evidence linking can achieve citation rates in the high ranges. The barrier is not suppression – it is credential and evidence requirements that most smaller sites cannot meet without fundamentally restructuring how they produce content.
Boundary condition: YMYL trigger rates and suppression boundaries are actively shifting. The September 2025 Quality Rater Guidelines update expanded YMYL scope; future updates may expand or contract specific suppression categories. Hard suppression boundaries – crisis mental health, election queries, medication dosage specifics – have shown stability, but partial suppression categories shift with algorithm updates. Verify current category-level trigger rates against SE Ranking or BrightEdge YMYL studies before forecasting citation potential.
Sources
- Position Digital – Ai Seo Statistics
- SE Ranking – Ai Overviews And Ymyl Topics Research
- BrightEdge – Healthcare Ai Evolution Google 2023 2025
- UpGrowth – Ymyl Playbook Healthcare Brands Win Ai Search Trust
- Search Engine Journal – Can You Use Ai To Write For Ymyl
- SEOZoom – Google Search Quality Rater Guidelines
- Originality.AI – Ai Overview Ai Citations Study
- Local SEO Guide – How To Rank At The Top Of Google Ai Overviews
- Hashmeta – The Eeat Framework For Ai Search
- UpGrowth – Google Ai Overviews Healthcare Traffic Data