search
arrow_forwardContact us

The AI Paradox: High-quality, human-created content leads to better visibility in AI search

Image of copywriter Petter Fløttum-Angeltveit in AG - AlfGundersen

Petter Fløttum-Angeltveit

Are you planning to rank high in Google searches with a website full of thin content written solely to include the right keywords? Well, Google says, just like Bart Simpson: E-E-A-T my shorts!

So far this year, Google has rolled out two major updates that continue to steer its search engine in a clear direction: away from clickbait and toward relevance, depth, and expertise. Pure AI services like Claude and ChatGPT prioritize these same values.

The goal for all of them is for you and me to get the best possible answers to our questions. To achieve this, the answers must be based on content that is as relevant, accurate, up-to-date, and credible as possible. But how do Google and AI services determine what meets the criteria?

Let them E-E-A-T cake!

No, Marie Antoinette probably never uttered those famous words. But just like a hungry French populace, search engines and AI systems love E-E-A-T, as it’s known in plain English.

E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trustworthiness. In other words, experience, expertise, authority, and trustworthiness. It is Google’s framework for assessing content quality. All major AI services use similar quality signals when evaluating content.

In a nutshell, E-E-A-T is about whether the content:

  • was created by someone with personal experience on the subject
  • is produced by someone with documented expertise on the subject
  • The website and the author are credible in the field
  • The website is secure, transparent, and accurate

Here you can read Google’s rather extensive checklist of questions you can ask about your own content to see if it meets the E-E-A-T criteria.

The framework is far from new, but what is new is that it is now being applied in far more fields than before and is being enforced more strictly. The goal is to ensure that we receive content that is as high-quality, relevant, accurate, and helpful as possible when we ask Google or AI solutions for help.

And how do you do that? By creating content such as:

  • Is signed by a named person
  • Based on real-world experience
  • Includes specific figures, examples, and sources
  • Takes a clear stance
  • Sticks to a single theme or subject area over time
  • Is technically sound (you can read more about what that meanshere and here )
  • Updated regularly

It is important to note that expertise is evaluated at the topic level. This means that websites with extensive and credible content on a given topic will rank higher in searches for that topic than individual articles on websites that otherwise have little content on that topic.

Want to learn more about how to improve your visibility in AI search? Check out these articles we’ve written:

Search optimization for AI: How to be visible in the new AI world

Schema Markup: What Is It and Why Is It Important for SEO and AI?

Punishes poor AI, not necessarily AI itself

Although Google and AI services such as Claude, ChatGPT, Perplexity, and Gemini evaluate content based on the same principles, each has its own weightings and nuances. As a result, the same search will yield different answers and different sources across these various services.

Peec AI analyzed 30 million citations across ChatGPT, Google AI Mode, Gemini, Perplexity, and AI Overviews. Reddit emerged as by far the most cited source, followed by YouTube. For B2B-related topics, LinkedIn is the most cited, according to Profound data.

Perhaps the most striking finding from the studies mentioned is that all the websites that frequently rank highly feature authentic, human-created content with clearly identified authors. Another analysis of Google’s March update shows that primary sources are now weighted more heavily at the expense of secondary sources. In other words: Do you want to be highlighted by AI? Have content created by humans, for humans. Not by AI, for AI.

At the same time, it is important to emphasize that Google does not penalize AI-generated content per se. John Mueller from Google stated in November 2025 that what matters is not whether the content was created by AI or humans, but whether the page adds value.

They aren’t punishing AI; they’re punishing“AI trash.” Or AI crap, if you will. In other words, quick and cheap content without editorial judgment, fact-checking, or any real thought.

Above you see an example of Google's AI summary (AIO). Illustration: Screenshot from Google.

The hardworking were rewarded for their efforts

Following Google’s March update, many professionals working in digital visibility and search engine optimization were on the edge of their seats. The update had a major impact on search results, and nearly 80 percent of the top three results changed, according to Search Engine Land.

The update had varying effects on our clients. In Google Search Console, we saw a clear trend: websites with high-quality, well-researched, and customer-centric content were rewarded with improved visibility in Google. And those that hadn’t focused on such content? Well, they were penalized.

Do you want to stand out in the new online search landscape? Then it’s not enough to just follow E-E-A-T, Pray, Love. You actually need to set aside time and resources to produce content that demonstrates insight, credibility, and expertise.

Frequently Asked Questions About AI Search

It remains difficult to measure visibility in AI search engines other than Google, because none of the others publish actual figures. There are several services that claim they can measure AI visibility, but they rely on mass prompting. This means they ask AI services large numbers of questions and then measure the visibility in the answers they receive. This provides a rough estimate, but the data is far from as reliable as what Google provides.

E-E-A-T stands for Experience, Expertise, Authoritativeness, and Trustworthiness. This is the framework Google uses when evaluating whether a website is good or not. E-E-A-T is not a direct ranking factor, but the signals on which the framework is based have a significant impact on your visibility in both Google Search and AI Search.

A website is rarely equally good at everything. Google and AI services therefore assess expertise topic by topic, not for the website as a whole. A site that has published a lot of credible content on a particular topic over time will rank higher in searches for that specific topic than a much larger site that just happens to have a single article on it.

Because these platforms are full of real people sharing real experiences, knowledge, and discussions. AI services interpret this as credible sources because the content is written by people with actual firsthand knowledge—exactly what the E-E-A-T framework prioritizes. For B2B topics, LinkedIn is particularly important, while Reddit dominates for consumer-related questions, even though ChatGPT has scaled back how much it highlights Reddit.

Because the author of the article loves wordplay.

GEO (Generative Engine Optimization) is the most commonly used term to describe how to optimize web content so that it is found—and actually used—by generative AI solutions such as ChatGPT and Google’s AI summaries. The phenomenon is also referred to by other acronyms that essentially cover much of the same ground, including AEO (Answer Engine Optimization), SOFA (Search Optimization for AI), LLMO (Large Language Model Optimization), LLM SEO (Large Language Model Search Engine Optimization), and AI SEO (Artificial Intelligence Search Engine Optimization).

An LLM (Large Language Model) is a language model trained on huge amounts of text to understand, summarize and create language similar to how humans write and speak. An LLM is the engine that drives generative AI services, enabling them to understand questions and produce new content.

Sources:

Disclaimer: Yes, I used both Claude and Google while working on this article, but it was written by a real person (me) for a real person (you).

← Previous article Schema Markup: What Is It and Why Is It Important for SEO and AI?