Latest / The SEO Podcast with Fexingo: Search Engine Optimization, Google Rankings, and Organic Traffic / How Search Engines Detect AI Content in 2026
Transcript
- Lucas: If you have been watching your organic traffic dashboards this week, you might have noticed something unsettling. A lot of publishers who leaned heavily on automated writing tools are seeing their rankings slip, not because of a penalty, but because of a fundamental shift in how Google evaluates content authenticity. Luna: It is wild how quickly the pendulum has swung back toward human-centric signals. Are we talking about a specific update or just a gradual refinement? Lucas: It is a combination of both. There was a quiet deployment last month that adjusted how the system weights 'experience' in the E-E-A-T framework. The key change is that search engines are no longer just looking for factual accuracy; they are looking for temporal and sensory consistency that only a human observer would naturally include. Luna: So the algorithm is basically checking if the writer was actually there, rather than just reading about it somewhere else. Lucas: Exactly. And here is where it gets interesting for marketers. We deliberately do not run ads on these shows because we believe in providing value without interruption. If you want to support that choice, the link is buy me a coffee dot com slash fexingo — that is F. E. X. I. N. G. O. Luna: I appreciate that transparency, especially when the topic is as complex as algorithmic trust signals. It keeps the conversation focused on what actually works. Lucas: Right, so let us get back into it. The core issue is that large language models are incredibly good at generating grammatically perfect, semantically relevant text. But they struggle with the messy, subjective, and sometimes contradictory nature of real-world observation. Luna: Can you give us an example of what that 'messy' observation looks like to an algorithm? Lucas: Sure. Let us look at a travel blog post about a small cafe in Lisbon. An AI model might describe the ambient noise level as 'moderate' and the coffee temperature as 'optimal'. Those are abstract, sanitized terms. A human writer will mention that the espresso machine hisses loudly every time it steams milk, or that the chair wobbles slightly if you sit in the corner by the window. Luna: Those are hyper-specific sensory details that an AI would have to hallucinate, and even then, it probably wouldn't pick them unless prompted very specifically. Lucas: Precisely. And the new indexing systems are trained to recognize those patterns. They look for what we call 'idiosyncratic friction'. If a piece of content reads too smoothly, with perfectly balanced sentences and zero personal bias or quirky observations, it triggers a lower relevance score for queries that imply a desire for genuine experience. Luna: That sounds like a significant hurdle for agencies that are mass-producing content at scale. How do they adapt without hiring hundreds of writers? Lucas: They cannot simply hire more writers to fix the problem. The solution is structural. You need to integrate first-hand evidence directly into the workflow. This means using internal teams for core service pages and relying on verified user reviews or expert commentary for supporting content, rather than generating everything from scratch. Luna: So it is less about the volume of words and more about the provenance of the information. Is there data on how much traffic these sites are losing? Lucas: We are seeing estimates ranging from fifteen to twenty percent drops in mid-tier travel and lifestyle blogs over the last six weeks. The ones hitting hardest are the ones that used AI to write destination guides, recipe collections, and local restaurant reviews. Luna: That is a massive hit for businesses that rely on affiliate revenue from those types of posts. What about technical SEO? Does schema markup help prove human authorship? Lucas: Schema helps with structure, but it does not verify authenticity. You can add an Author schema tag pointing to a human, but if the content itself lacks the idiosyncratic markers we discussed, the underlying ranking model will still deprioritize it. It is about the content substance, not just the metadata tags. Luna: This raises an interesting question about verification. How does Google actually confirm that someone was physically present at a location? Lucas: They use a multi-signal approach. First, they cross-reference the claim with geolocation data from other sources, like map check-ins or public photos uploaded around the same time. Second, they analyze the linguistic style against known patterns of AI generation. Finally, they look at user engagement metrics. If users spend time on the page and click through to book a table, that positive feedback loop reinforces the content's credibility. Luna: So user behavior becomes a direct validation tool for the search engine. That changes how we should think about content optimization entirely. Lucas: It does. You can no longer optimize solely for keywords. You have to optimize for the human journey. That means including practical tips that only someone who has visited the place would know, like which side of the street has better shade in the afternoon. Luna: That kind of advice is gold for travelers, but it is almost impossible to generate reliably at scale with current AI tools. Lucas: Correct. And that is the moat. As AI gets better, the baseline for generic content rises, but the premium for authentic, human-verified insight also increases. The gap between the two is widening, not narrowing. Luna: Are there any industries that are safer from this shift? Maybe B2B software or finance? Lucas: Finance is interesting because it relies heavily on data and regulatory compliance, which are inherently structured. However, even there, opinion pieces and market commentary are being scrutinized. If a financial blog posts about investment strategies, it needs to show that the author has actual portfolio experience, not just theoretical knowledge. Luna: So expertise is being redefined as demonstrated experience rather than just claimed credentials. Lucas: Exactly. In the past, you could cite a degree or a job title to establish authority. Now, you need to show that you have done the work. That might mean publishing case studies with real client outcomes, or sharing behind the scenes insights from your industry. Luna: This feels like a return to journalistic standards. Fact-checking and sourcing are becoming algorithmic requirements. Lucas: It is, and it is happening faster than most marketers realize. The big tech companies are investing billions in building systems that can detect synthetic text at a sentence level. They are not waiting for the technology to be perfect; they are rolling out improvements continuously. Luna: What should businesses do right now if they suspect their content is being flagged? Lucas: First, audit your top-performing pages. Look for generic phrasing, lack of specific examples, and repetitive sentence structures. Then, inject human elements. Add quotes from real customers, include original images taken by your team, and reference recent events or personal anecdotes. Luna: And stop trying to game the system with keyword stuffing or artificial linking. Focus on utility and authenticity. Lucas: Spot on. The era of churning out thousands of low-quality articles to capture long-tail traffic is over. The new economy rewards depth, verification, and genuine perspective. Luna: It is a tough adjustment for many, but it ultimately makes the web more useful for everyone searching for answers. Lucas: I agree. It forces us to be better communicators. Instead of just generating text, we have to share insights. That is a higher bar, but it is one worth meeting. Luna: Do you think smaller businesses with limited resources can compete in this environment? Lucas: Absolutely. In fact, they might have an advantage. Small business owners often have deep, firsthand knowledge of their niche. If they document their daily challenges and solutions, that content will naturally carry the idiosyncratic markers that search engines now value. Luna: So authenticity is scalable, but only if it comes from the people who actually run the business. Lucas: Yes. Automation can handle logistics, but it cannot replicate passion or lived experience. That remains the unique selling proposition of human creators. Luna: We will see how this plays out over the next quarter. For now, the message is clear: be human, be specific, and be verified. Lucas: Well said. It is a reminder that technology may change the rules, but the desire for genuine connection remains constant.