September 2025

Building a presence on these platforms are crucial to having an AI presence Generative AI platforms like ChatGPT and Google’s Gemini are no longer novelties; they are the new information gatekeepers. When asking a question, unlike Google results, they don’t list links—they provide a single answer synthesized from sources they deem credible.  For businesses, content creators, or individuals concerned with online reputation, this raises a critical question: where exactly are they getting this information? And how can I use this to build an AI presence? Understanding which websites these AI models trust and cite is the first step in a new digital strategy of Generative AI Optimization (GAIO) and GenAI Reputation Management. To be visible in the AI-driven answers, you need to know which sources are shaping deep learning models.  Recover Reputation analyzed multiple large-scale studies and conducted direct research to deconstruct the information ecosystems of the two biggest Large Language Models. Knowing these platforms are crucial to building an AI presence. Here are the definitive lists of the websites that ChatGPT and Gemini rely on the most. The ChatGPT Canon: Authority and Community Rule ChatGPT’s sourcing strategy is built on a core “canon” of trusted domains. It has a clear preference for two types of content: authoritative, encyclopedic knowledge and vast, community-vetted conversations. This is supplemented by established media outlets and specialized review sites for consumer-related questions. Across the board, two giants stand out: Wikipedia for factual information (cited in 7.8% to 15% of cases) and Reddit for real-world experience (cited anywhere from 1.8% to a staggering 29.4% of the time, depending on the query type). This reliance is so significant that it’s clear these two platforms form the foundational pillars of ChatGPT’s knowledge base.   Here is a consolidated ranking of the top 20 domains most frequently cited by ChatGPT, along with their share of citations as found in major studies. Rank Domain Primary Category Share of Citations (%) Source Study 1 reddit.com Conversational UGC 1.8% – 29.4% Ahrefs, Profound 2 wikipedia.org Encyclopedic UGC 7.8% – 15.0% Ahrefs, Profound 3 forbes.com News / Media 1.1% – 6.7% Ahrefs, Profound, Wellows 4 businessinsider.com News / Media 0.8% – 1.3% Ahrefs, Profound 5 techradar.com Tech Review 0.9% – 11.8% Profound, Wellows 6 amazon.com E-commerce ~3.4% Ahrefs 7 nypost.com News / Media 0.7% – 1.0% Ahrefs, Profound 8 g2.com Software Review ~1.1% Profound 9 nerdwallet.com Finance ~0.8% Profound 10 thespruce.com Lifestyle / Home ~1.3% Ahrefs 11 cnet.com Tech Review ~8.8% Wellows 12 pcmag.com Tech Review ~7.0% Wellows 13 wired.com Tech / Media ~1.0% Ahrefs 14 reuters.com News / Media ~0.6% Profound 15 tomsguide.com Tech Review ~4.6% Wellows 16 bhg.com Lifestyle / Home ~1.0% Ahrefs 17 people.com Entertainment / Media ~1.0% Ahrefs 18 techcrunch.com Tech / Media ~4.0% Wellows 19 hbr.org Business / Media ~2.8% Wellows 20 openai.com Corporate / Tech ~2.8% Wellows Gemini’s Playbook: Context is Everything Google’s Gemini operates slightly differently. Instead of relying on a fixed set of top domains, it acts as a “balanced synthesizer,” dynamically choosing its sources based on the specific topic of the query. This makes its citation patterns more diverse and highly specialized.   One of Gemini’s biggest advantages is its deep integration with its own ecosystem, especially YouTube, which accounts for approximately 3% of its citations in some studies. For health queries, it shows a unique preference for government and NGO sources, citing them nearly 25% of the time.   Because Gemini’s sources change dramatically depending on the topic, we’ve broken down the top domains by category. Top 20 Cited Domains for General Queries (Google AI Mode) For broad, everyday questions, Gemini (powering Google’s AI Mode) pulls from a wide range of user-generated content, reference sites, and major online platforms.   en.wikipedia.org (12.0% share)   www.youtube.com (1.8% – 10% share)   blog.google www.reddit.com (2.2% – 14% share)   www.google.com (7.4% share)   www.amazon.com www.quora.com (1.5% share)   www.facebook.com m.yelp.com www.instagram.com www.imdb.com www.tripadvisor.com www.linkedin.com (1.3% share)   www.mapquest.com www.walmart.com www.britannica.com www.healthline.com www.yahoo.com www.ebay.com my.clevelandclinic.org Top Cited Domains for Health & Medicine When it comes to health, Gemini shows a strong preference for official, institutional, and highly authoritative medical sources over general media.   pmc.ncbi.nlm.nih.gov (PubMed Central) (~7.0% share) my.clevelandclinic.org (~3.2% share) www.mayoclinic.org (~3.0% share) www.ncbi.nlm.nih.gov (National Center for Biotechnology Information) (~2.7% share) www.sciencedirect.com (~1.7% share) www.healthline.com www.webmd.com www.medicalnewstoday.com www.verywellhealth.com www.goodrx.com medlineplus.gov www.drugs.com www.cdc.gov (Centers for Disease Control and Prevention) Top Cited Domains for Automotive For car and auto insurance queries, Gemini leans on a mix of specialized review sites, industry authorities, and major media outlets. bankrate.com (6.7% share) thezebra.com (7.2% share) nerdwallet.com edmunds.com kbb.com (Kelley Blue Book) caranddriver.com cars.usnews.com www.cars.com forbes.com en.wikipedia.org reddit.com youtube.com Top 20 Cited Domains for B2B Tech For business-to-business technology questions, Gemini shifts its focus to company blogs, niche industry publications, and professional platforms.   Company Websites/Blogs (~17% share) Niche B2B Publications (e.g., TechTarget) Mainstream News (~10% share) linkedin.com (~2% share) Analyst Reports (e.g., Gartner) forbes.com businessinsider.com pcmag.com cnet.com techradar.com tomsguide.com techcrunch.com hbr.org (Harvard Business Review) zapier.com (Blog) medium.com www.nytimes.com www.cnbc.com play.google.com apps.apple.com www.investopedia.com What Does This Mean for You? These lists reveal a clear roadmap for anyone looking to build authority, visibility and a reputation in the age of AI. The models are designed to prioritize signals of trust and expertise.   Authority is Paramount: High-authority domains like Wikipedia, Forbes, and major health institutions are consistently favored. Building genuine credibility in your niche is more important than ever. User-Generated Content is King: Platforms like Reddit and YouTube are not just social networks; they are massive repositories of human experience that AI models rely on heavily. Authentic participation in these communities is extremely crucial. Content Must Be Contextual: For Gemini, in particular, the best source depends on the topic. Your content strategy must be tailored to your specific industry, whether that means creating in-depth health guides, authoritative financial reviews, or engaging B2B tech videos. As AI continues to evolve, the websites it trusts will shape what the world knows. By understanding these preferences, you can position your content to be a source of truth for both humans and the machines that guide

Firm’s Patent-Pending Synergistic Reputation Repair™ is the First Reputation Management Solution for the New Problem of AI Misinformation. NEW YORK, NY – September 4, 2025 – Recover Reputation, an online reputation management firm, today announced its patent-pending Synergistic Reputation Repair™, a new solution designed specifically to correct inaccurate and damaging answers about businesses and professionals appearing in AI chatbots like OpenAI’s ChatGPT and Google’s Gemini. The launch provides a direct answer to the new and urgent problem of AI-generated misinformation, where incorrect answers from chatbots can damage a company’s brand, mislead customers, and create significant business risks.  Synergistic Reputation Repair™ is the first systematic framework designed to repair AI-generated misinformation at its source. It moves beyond outdated online reputation management and SEO tactics, which are ineffective against the synthesized, authoritative-sounding narratives produced by Large Language Models (LLMs). “AI has become the new front page for everyone—from businesses and professionals to underrepresented groups who are often disproportionately harmed by algorithmic bias. But it frequently gets the facts wrong, and until now, there hasn’t been a clear way to fix it,” said Steven W. Giovinco, founder of Recover Reputation. “Our solution provides the first direct, systematic process for correcting the record. We are committed to ensuring that everyone has the right to a fair and accurate digital representation in the age of AI.” The proprietary, three-part system works synergistically to deliver durable results: Proactive Content Strategy: Creates and promotes a portfolio of accurate, authoritative content across the web. This provides AI models like ChatGPT and Gemini with a reliable foundation of factual information to draw from when generating answers about a person, business or group. Direct AI Correction: Engages directly with AI platforms to correct false and misleading statements. Using AI’s feedback systems, this process systematically flags inaccuracies and reinforces the correct information, making the corrections more effective and durable. Long-Term Reputational Shielding: Develops a structured, high-quality dataset of verified information. This serves as a long-term asset to fortify their reputation and protect against future AI-generated inaccuracies. Recover Reputation is one of the first firms to tackle this emerging threat and the only one with a patent-pending, integrated system designed for the unique challenges of the AI era. Based on documented case studies, comprehensive campaigns are designed to achieve significant and lasting transformations within a six-month timeframe.   About Recover Reputation Recover Reputation is a New York-based online reputation management firm specializing in correcting complex misinformation in AI platforms. Founded by 30-year technology veteran Steven W. Giovinco, the company is the inventor of the patent-pending Synergistic Reputation Repair™ framework, the only solution engineered to combat the new and complex threats of the AI era and promote digital equity.   Media Contact: Steven W. Giovinco Founder, Recover Reputation steve@recoverreputation.com +1 347-559-4952 www.recoverreputation.com ###