The EPR Authority Index: 100 High-Authority Domains That Shape SEO, Reputation, and AI Visibility

A curated authority map across thirteen categories — government, knowledge, academia, research, news, technology, professional networks, commerce, community, data, legal, business, and vertical industry authorities. Grouped, not ranked.

By Ronn Torossian

Authority on the internet is no longer concentrated in four places. The old framework — government, academia, news, journals — still matters, but it now sits inside a wider citation ecosystem. AI engines learn from and cite platforms that dominate knowledge creation, software development, professional identity, commerce, and community consensus — and they treat the leaders in each as authoritative.

ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews all weight source authority before they cite a fact. A Reuters byline still matters. So does a GitHub repository, a Reddit thread, a Mayo Clinic article, a SEC filing, and a peer-reviewed paper on PubMed. The hierarchy is wider than it used to be — and any serious conversation about authority has to account for the whole ecosystem.

This is the EPR Authority Index — a curated map of one hundred domains that meaningfully shape search, reputation, and AI visibility on the open internet. Organized into thirteen categories. Grouped, not ranked.

Why This Is A Map, Not A Ranking

A precise 1–100 ordering would be indefensible, for three reasons:

Page-level authority outweighs domain-level authority. A topically relevant Reuters article about your company outperforms a generic Harvard department page about something else. Google has been moving in this direction for a decade; the AI engines were built this way from the start.

Value depends on purpose. A backlink from CISA.gov outperforms Harvard for a cybersecurity company. A Reuters mention outperforms most government links for a PR campaign. A Mayo Clinic citation outperforms NYT for a hospital system. There is no single answer to "most valuable" — only "most valuable for what."

No public dataset can produce a defensible precise ordering. Inclusion in a curated authority map is defensible. Telling you GitHub is more valuable than Wikipedia is not.

Inclusion Criteria

Every domain on this list meets at least three of the following:

• Strong editorial or institutional gatekeeping — formal review, regulatory authority, peer review, or platform-scale curation.
• Demonstrable presence as a cited source across multiple AI engines (ChatGPT, Claude, Perplexity, Gemini, Google AI Overviews).
• Mature link or citation graph — typically a decade or more of inbound editorial links and references from other authoritative domains.
• Crawl access for major AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended) or, for paywalled domains, regular indexing through the editorial ecosystem.
• Category-defining authority within its vertical — the publication, platform, or institution that other publications cite by default.

The Index

Government & Public Institutions

Editorial gatekeeping is built into the institution itself. Among the hardest links to earn — and the most durable signal of authority for SEO and AI citation.

United States

• CDC.gov — Centers for Disease Control. Public-health authority of record.
• CISA.gov — Cybersecurity and Infrastructure Security Agency. The single most valuable cyber-PR backlink in the U.S.
• Congress.gov — Legislative record of the United States.
• FDA.gov — Pharma, medical-device, and food regulatory record.
• FederalReserve.gov — U.S. central bank. Speeches, research, monetary policy.
• NASA.gov — One of the most cited .gov domains in education and science.
• NIH.gov — Biomedical research gravitational center.
• SEC.gov — Every U.S. public-company filing.
• SupremeCourt.gov — Opinions and orders of the U.S. Supreme Court.
• WhiteHouse.gov — U.S. Executive Branch official communications.

United Kingdom

• BankOfEngland.co.uk — Monetary authority of the United Kingdom.
• GOV.uk — Unified U.K. government portal. The most-linked .gov.uk surface.
• Parliament.uk — House of Commons and Lords. Hansard, committees, evidence.

European Union

• EC.europa.eu — European Commission. The EU executive.
• ECB.europa.eu — European Central Bank. Euro-area monetary policy.

Supranational

• OECD.org — Default source for cross-country economic comparisons.
• UN.org — United Nations system.
• WHO.int — World Health Organization.
• WorldBank.org — Development economics data and reports.

Knowledge & Reference

The reference layer of the open internet. The platforms LLMs walk to when they need ground truth.

• Archive.org — Internet Archive. The web's memory. Citation source for primary documents and lost pages.
• Britannica.com — Encyclopædia Britannica. Editorial reference depth Wikipedia doesn't fully replicate.
• PubMed.ncbi.nlm.nih.gov — National Library of Medicine. The bibliography behind global health and biomedical research.
• Scholar.Google.com — Google Scholar. The default scholarly citation surface for the open web.
• Wikipedia.org — The most-cited domain on the open web. LLM training-data anchor.

Universities

The .edu and equivalent academic domains the AI engines weight more heavily than almost any commercial source. Page-level relevance matters: a citation from a research group active in your field outperforms a generic department link.

North America

• Berkeley.edu — Computer science, climate, public policy.
• Columbia.edu — Journalism, business, public health.
• Harvard.edu — The most-linked university domain in the academic graph.
• MIT.edu — Engineering, AI, economics.
• Princeton.edu — Origin point for foundational research, including GEO.
• Stanford.edu — Silicon Valley's academic home.
• Yale.edu — Law, medicine, humanities.

Europe

• Cam.ac.uk — Cambridge. Sciences and humanities anchor for the Commonwealth.
• ETHZ.ch — ETH Zürich. Continental Europe's technical research peak.
• LSE.ac.uk — London School of Economics.
• Ox.ac.uk — Oxford. The global academic authority outside the U.S. system.

Asia-Pacific

• NUS.edu.sg — National University of Singapore.
• Tsinghua.edu.cn — Tsinghua University. Engineering and technology authority in China.

Research Institutions & Scientific Journals

Peer-reviewed citation infrastructure plus the great research institutions outside the university system.

Journals

• ArXiv.org — Preprint server. The fastest-citation surface in physics, math, and AI.
• Cell.com — Life sciences citation peak.
• Nature.com — The world's most-cited science journal.
• NEJM.org — New England Journal of Medicine. The reference standard in clinical medicine.
• Science.org — AAAS. Co-equal to Nature in citation gravity.
• TheLancet.com — Global health and medicine reference.

Research Institutions

• CAS.cn — Chinese Academy of Sciences. Largest scientific research organization in the world by headcount.
• CERN — European Organization for Nuclear Research. Reference for fundamental physics.
• MPG.de — Max Planck Society. Germany's research network across 86 institutes.
• RoyalSociety.org — Founded 1660. Among the oldest scientific bodies on the internet.

News & Editorial

Reuters, AP, and AFP function as wire-level ground truth — syndicated, cited, and rewritten across thousands of downstream outlets, which makes their effective citation graph larger than any single national newspaper.

Global wires

• AFP.com — Agence France-Presse. The third major global wire.
• AP.org — Associated Press. The other wire LLMs treat as ground truth.
• Bloomberg.com — Financial data plus journalism. Terminal-grade authority.
• Reuters.com — Among the most-syndicated news sources on the internet.

Tier-one papers

• BBC.com — Global broadcaster. The highest-trust international news domain.
• Economist.com — Long-form policy and business analysis.
• FT.com — Financial Times.
• NYTimes.com — Paper of record.
• TheGuardian.com — U.K. paper of record outside the right-of-center mainstream.
• WashingtonPost.com — U.S. politics of record.
• WSJ.com — Default citation for U.S. market reporting.

Technology & Developer Infrastructure

The platforms where modern technical knowledge is created, versioned, and cited. LLMs were trained on this stack and continue to cite from it.

• ACM.org — Association for Computing Machinery.
• Developer.Mozilla.org — MDN Web Docs. Reference for web standards across every browser.
• GitHub.com — The world's source code. The single most-trained-on commercial domain.
• IEEE.org — Institute of Electrical and Electronics Engineers.
• IETF.org — Internet Engineering Task Force. Where RFCs live.
• Learn.Microsoft.com — Microsoft's developer documentation. Reference surface for .NET, Azure, Windows.
• StackOverflow.com — Q&A authority for programmers. Heavily cited training data.
• W3.org — World Wide Web Consortium.

Professional & Business Networks

Where professional identity, credentials, and corporate narratives are indexed.

• LinkedIn.com — Default professional identity layer of the internet. Increasingly cited by AI engines for biographical and employer context.

Commerce & Consumer Trust

Domains so foundational to consumer behavior that the AI engines treat them as ambient authority — product, brand, and trust signals propagate from here.

• Amazon.com — Default consumer product reference. Reviews, listings, and product copy seed AI consumer answers.
• Apple.com — Product authority across consumer hardware and software.
• BBB.org — Better Business Bureau. Trust and complaint signal in U.S. consumer queries.
• Google.com — Product, knowledge graph, and reference surface.
• Microsoft.com — Enterprise software authority.

Community & Discussion Platforms

Where consensus is formed in public. AI engines weight these platforms heavily because they capture how real people actually talk about brands, products, and ideas.

• Reddit.com — Now formally licensed to Google for AI training. The single highest-weighted community-discussion source across modern LLMs.
• YouTube.com — World's largest video reference surface. Transcripts feed AI answers across consumer and how-to queries.

Data & Market Intelligence

The paywalled and semi-paywalled data layer that financial, business, and analyst-grade content rests on. Disproportionate citation weight inside enterprise queries.

• Crunchbase.com — Default startup and funding-round reference.
• Moodys.com — Credit ratings authority.
• PitchBook.com — Private-market data authority used by venture and PE.
• SPGlobal.com — S&P Global. Indices, ratings, market intelligence.
• Statista.com — Stat-citation default across business writing on the open web.

Legal & Regulatory Intelligence

The reference layer behind legal writing, regulatory filings, and case research.

• CourtListener.org — Free Law Project. Open-access case law.
• Justia.com — Free legal information and case law.
• LexisNexis.com — Legal, regulatory, and news intelligence platform.
• Westlaw — Thomson Reuters Westlaw. Legal research standard for U.S. practice.

Business & Management Authority

The institutional voices buyers and boards already trust. A backlink from any of these is a credential as much as a link.

• Bain.com — Bain & Company. Private equity and consumer strategy depth.
• BCG.com — Boston Consulting Group.
• Forrester.com — Tech analyst with deep marketing and CX coverage.
• Gartner.com — IT and software analyst authority. Magic Quadrant citations are commercial gold.
• HBR.org — Harvard Business Review. The most-cited business publication on the open web.
• McKinsey.com — Firm of record for C-suite citation.

Industry Authorities by Vertical

Inside any given vertical, a backlink from the category authority can outperform a generic link from a tier-one news brand. These are the domains that define their industries.

Healthcare

• ClevelandClinic.org — Cleveland Clinic. Consumer-health authority and treatment reference.
• MayoClinic.org — Mayo Clinic. The most cited consumer-health domain in AI answers.

Cybersecurity

• MITRE.org — MITRE ATT&CK framework. The reference taxonomy for adversary behavior.
• OWASP.org — Open Web Application Security Project. Application-security standard.
• SANS.org — SANS Institute. Security training and research authority.

Defense, Policy & Think Tanks

• Brookings.edu — Brookings Institution. The most-cited U.S. policy think tank.
• CSIS.org — Center for Strategic and International Studies.
• RAND.org — RAND Corporation. Defense, security, and policy research.

Energy

• IEA.org — International Energy Agency. Global energy data and policy.

Technology Press

• TechCrunch.com — Startup and venture-capital reporting of record.
• TheVerge.com — Consumer-tech editorial authority.

How To Actually Earn These

A backlink or citation from a top-tier domain on this list is not a deliverable a junior link-builder produces. It is the byproduct of being the actual source on something the institution covers. Four routes work:

Original research. Government, academic, and tier-one news cite primary data they can attribute. Studies, indexes, and proprietary surveys are link magnets at the top of the pyramid.

Founder authority. A named executive with a published point of view gets quoted. Quotes become links.

Public records footprint. SEC filings, court records, patent filings, congressional and parliamentary testimony. If your work shows up in the official record, the linking follows.

Earned editorial. The classic press pitch — sharpened for the AI era, written for the prompt the buyer is going to type into ChatGPT next Tuesday.

Most agencies still sell link volume. The future belongs to the firms that earn a credential-grade citation, then make it work — across Google, across the AI engines, and across the buyer's path to the answer. 5W is built for that work.

About the Author