Author and trust signals for AI: what E-E-A-T actually is

Every guide to AI visibility ends up making the same promise: add a credentialed author, publish an About page, layer on trust signals, and ChatGPT and Perplexity will start citing you. The advice is half right. A real, named author with expertise that fits the topic, a way to reach the people behind the site, and substantive About and policy pages are all worth building.

The reason to build them matters, though. AI engines have no E-E-A-T dial to turn: E-E-A-T is the lens Google hands its human quality raters, and no controlled study has shown that a bio earns citations. Author and trust signals pay off because they earn human trust. The goal is accountability, not a checklist.

How AI engines actually see author and trust signals

There is no E-E-A-T score to optimize, in ChatGPT, Perplexity, Gemini, Google's AI Overviews or Microsoft Copilot. E-E-A-T (Experience, Expertise, Authoritativeness, Trust) comes from Google's Search Quality Rater Guidelines, the rubric human reviewers use to grade search results, and OpenAI, Perplexity and Microsoft document no E-E-A-T framework at all. Google is unusually direct about its status in its helpful content guidance: "While E-E-A-T itself isn't a specific ranking factor, using a mix of factors that can identify content with good E-E-A-T is useful." The raters who apply the framework do not move results either: "Search raters have no control over how pages rank. Rater data is not used directly in our ranking algorithms." And for AI surfaces specifically, Google's AI features documentation says: "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary."

The available engine guidance is much narrower than the usual E-E-A-T advice implies. Google says its AI surfaces rest on the same foundations as Search: "You can apply the same foundational SEO best practices for AI features as you do for Google Search overall." Its guidance also ranks the four E-E-A-T aspects unequally: "Of these aspects, trust is most important." Whether ChatGPT, Perplexity or Copilot work the same way is an inference; none of them document it. An author or organization named and described the same way everywhere should in principle be easier to attribute and cite than one whose identity shifts from page to page, but no engine documents that and no cited experiment measures it.

The evidence for anything stronger is thin. The largest controlled citation experiment to date, a 2026 study (Vishwakarma et al.) covering 252,000 trials across six LLMs, used "brand anonymization and counterbalanced source order to separate content effects from position bias"; the strongest drivers of citation were "topical relevance and list position", while "Completeness and trust cues add smaller gains". The GEO paper that defined generative engine optimization (Aggarwal et al., 2024) tests nine optimization methods, and none of them involves authorship, bylines or trust pages. So the case for the five moves below rests on human credibility, Google's stated foundations and an inferred corroboration effect, and they are worth making on those grounds. The five moves are: name a verifiable author, show expertise that fits the topic, be reachable with real trust pages, give those pages substance, and disclose money and AI.

Name a real, verifiable author

The bluntest signal is a named human being. A byline that reads admin, staff or editorial team gives a reader nothing to weigh and leaves nothing for anyone citing the page to attribute it to. Google's self-assessment questions for helpful content put it plainly: "Do bylines lead to further information about the author or authors involved...?" A name alone does not clear that bar; the byline should lead somewhere.

Verifiable means three things in practice. First, the author is visible on the page itself, not only buried in metadata. Second, the visible name and any structured data agree: a page that says "Jane Doe" in the byline but carries "Site Admin" in its JSON-LD author field signals templated metadata, and a reader who spots the mismatch has less reason to trust either value. Third, the byline links to a substantive author page: a bio with the person's background, their other writing on the site, and, where they exist, links to independent profiles elsewhere. Marking the author up with Person schema that points at those profiles is optional, but it expresses the same identity in machine-readable form: the same person, described the same way, in more than one independent place. Whether that changes citation selection has not been measured.

Show expertise that fits the topic

An author bio earns its keep by establishing expertise for this page's subject, not generic authority. Google's rater guidelines make the point with a scenario: for advice on rewiring a house, you would trust a skilled electrician over an enthusiast of antique homes who knows nothing about electrical work. A credential is only meaningful relative to the topic it sits next to.

The bar is also lower than most E-E-A-T guides admit. Google's self-assessment asks: "Is this content written or reviewed by an expert or enthusiast who demonstrably knows the topic well?" The words "or enthusiast" are doing deliberate work. A recipe page is well served by someone who demonstrably cooks. A travel guide is well served by someone who has been there. A woodworking site is well served by years of documented projects. Inventing or padding credentials is unnecessary and usually obvious to readers. The place to raise the bar is where the stakes rise: medical, legal and financial topics, what the rater guidelines call Your Money or Your Life (YMYL) pages, are where formal, verifiable qualifications matter most.

There is no expertise score behind any of this. Whether a bio fits a topic is a judgment made by humans, and by systems tuned on human judgments. The aim is a bio a skeptical reader would accept as relevant to the page it sits on.

Be reachable and carry real trust pages

Author credibility also rests on the site behind the byline, and two baseline signals show that a real organization runs it. The first is a way to reach the people who run it: a contact page, or a concrete route such as an email address, a phone number or a postal address. The second is the set of institutional pages readers expect: an About page, a privacy policy, terms, a copyright line. Google's rater guidance starts its trust read at the About page, and raters are told to look for who is responsible for the site.

The evidence here is about human trust, and it predates AI answer engines entirely. BJ Fogg's Stanford web-credibility research (CHI 2001) found that "real-world feel" markers, a physical address, a phone number, an email address, were among the strongest drivers of perceived credibility. Stanford's Web Credibility guidelines distilled the same research into direct advice: "Show that there's a real organization behind your site" and "Make it easy to contact you." This evidence supports the value of these details for human readers, but it does not show that they affect AI citations.

These signals live at the site level, usually in a shared footer, so they are earned once and inherited everywhere. An individual article does not need its own contact block; it needs to sit on a site where the contact route and the policies exist and resolve.

Make your trust pages substantive

Presence is not substance. An About page that says "coming soon", a privacy policy pasted from a generator with the placeholders still in it, a contact page whose form goes nowhere: these stubs fail to provide the ownership and contact information raters look for. The rater guidelines are explicit about where this bites hardest: on pages that handle money or other YMYL topics, "an unsatisfying amount" of customer-service and contact information, or of "information about who is responsible" for the site, earns a Low quality rating.

The bar scales with the page type. An online store or a medical reference needs real depth: who owns the site, how to reach them, what happens to a customer's money and data. A humor site, a personal blog or a forum legitimately runs lean, and the same rater guidelines say that on forums and social platforms "a username is an acceptable identifier". Anonymity is not a defect where the page type warrants it: a pseudonymous personal site with years of consistent writing carries its own kind of accountability, and nobody expects a home address on it.

The test is whether the pages answer the questions a skeptical visitor would actually ask, not whether a policy would survive legal review.

Disclose money and AI

Transparency is how a page earns the Trust in E-E-A-T, and two disclosures carry real weight. Where a page earns from affiliate or sponsored links, saying so is not a nicety: in the United States the FTC's Endorsement Guides require a material connection to be disclosed "clearly and conspicuously", and in the United Kingdom the ASA does the same. The FTC also settles a placement question most guides leave open: a disclosure that sits only on an About page, or at the end of a post, is likely to be missed. Where content is materially AI-generated, Google's own self-assessment asks the question directly: "Is the use of automation, including AI-generation, self-evident to visitors through disclosures or in other ways?"

Scope matters here. Most pages have nothing to disclose: a plain recipe, a news report or a documentation page carries no affiliate relationship and needs no disclosure block, and treating every page without one as deficient would be wrong. The signal applies only where there is a real relationship or real automation to disclose. Where it does apply, a clear disclosure reads as editorial maturity rather than an admission. A site that tells readers how it earns money and how its content is produced is easier for a human to trust, and it gives anyone quoting it, human or machine, the facts to describe it correctly.

Frequently asked questions

Does adding an author bio get my page cited by AI?

Not as a measured lever. Google says E-E-A-T is not a specific ranking factor and its raters do not affect ranking, and no controlled study shows a bio earns citations; the largest cited experiment stripped brand and publisher names as a control and never tested author identity at all. A real, verifiable author still helps: it earns human trust, and it plausibly makes you easier to corroborate, though no study has measured that.

Do I need formal credentials to be cited?

Only where the topic demands it. Google's stated bar is an expert or enthusiast who demonstrably knows the topic well, so hands-on experience is enough for a recipe, hobby or travel page. Formal credentials matter most on YMYL topics (medical, legal, financial), where the trust bar is highest.

Is anonymous content penalized in AI search?

Not universally. The trust bar scales by page type: stores and pages about money or health need clear ownership, while personal sites, forums and user-generated content can legitimately use an alias. Google's rater guidance accepts a username as an identifier in those contexts, so anonymity there is not a defect.

Do About and privacy pages help if they are just placeholders?

No. A linked but empty About page or a boilerplate policy is a stub. Google's rater guidance treats an unsatisfying amount of ownership or customer-service information on money and YMYL pages as a Low quality signal. Write real pages that say who you are and how to reach you, not shells.

Is there an E-E-A-T score I can optimize?

No. Google is explicit that E-E-A-T is not a specific ranking factor, and no score exists to optimize. Treat it as a lens for building genuine credibility: a real author, real accountability behind the site, and honest disclosure where money or automation is involved, rather than a checklist to game.