[Host] Welcome to the A.E.O. Engine AI Search Show, the A.E.O. podcast for brands looking to earn citations in ChatGPT, Gemini, and Perplexity. I’m your host, Vijay Jacob, Founder and CEO of A.E.O. Engine. Today, we're looking at a shift that's fundamentally changing how we consume information online, especially something as critical as our health. We’ve got Marcus Reid, an industry analyst and former founder, joining us to break it down.
[Guest] Hey everyone, great to be here.
[Host] Marcus, let's start with a feeling many people have experienced lately. Picture this: you're scrolling through TikTok, maybe you’ve got a weird symptom, or you're just curious about a condition. You used to find these great, visually clear infographics made by actual healthcare professionals or reputable organizations. They were easy to digest, trustworthy. But lately, when you search for health information, especially on platforms like Google or even TikTok, you’re seeing something different. Instead of curated content, you're getting AI-generated summaries, sometimes even AI-created "doctor" avatars spouting advice. It’s frustrating, right? That feeling of, 'Wait, where did the reliable info go?'
[Guest] . I’ve seen that frustration echoed. The original prompt that sparked this discussion, "and then it got replaced with fricking ai overviews ☹️ #medicaltiktok," perfectly captures that user sentiment. It’s that moment of realizing a familiar, trusted pathway to information has been supplanted by something automated, something less tangible, and potentially less reliable.
[Host] Exactly. And that feeling, that disruption of a trusted experience, has a name. We’re talking about the phenomenon of AI Overviews and AI-generated content replacing human-curated medical information. It’s a significant shift, and it’s raising some serious questions about public health and trust online.
[Guest] Right. So, let's break down what this actually is. What are we seeing when we talk about these AI overviews and AI doctor personas?
[Host] What we're seeing are two primary manifestations. First, on platforms like Google Search, familiar search result pages are now frequently topped with AI Overviews. These are generative AI summaries that aim to give you a quick answer. But the problem is, they’re often pulling from web content in ways that can lead to significant inaccuracies. The Guardian, in a 2026 investigation, found these overviews frequently serving up dangerous health advice. Think recommending harmful treatments or misrepresenting medical consensus.
[Guest] So, it’s not just a different format; it’s actively misleading information presented as fact, right at the top of search.
[Host] Precisely. And then on platforms like TikTok, we’re seeing AI-generated "doctor" videos. These use deepfake or synthetic avatars, often with AI-generated voices, posing as medical professionals. They use phrases designed to sound authoritative, like one documented instance of an avatar claiming "13 years as a coochie doctor," which is… certainly a way to try and build credibility. Media Matters documented dozens of these accounts, and some of these videos were averaging 2.5 million views. That’s a massive reach for fabricated health tips or supplement promotions.
[Guest] And this isn't just theoretical. The research shows these AI platforms can even invent diseases, right? I recall reading about an experiment where a fake disease was created and then propagated by AI.
[Host] That's the third key manifestation: fabricated diseases and statistics. A research team deliberately invented a fake disease, let's call it "Bixonimania" for the sake of argument, and posted fabricated information online. Within days, multiple AI platforms – including Bing Copilot, Google Gemini, Perplexity AI, and ChatGPT – picked up this fake disease. Gemini might advise a user to see an ophthalmologist for symptoms related to it, Perplexity might invent a prevalence figure, and ChatGPT could directly attribute a user's symptoms to this non-existent condition. The truly alarming part? This fabricated information was subsequently cited in peer-reviewed medical literature, creating a dangerous feedback loop.
[Guest] That’s… deeply concerning. It’s one thing for an AI to get a fact wrong, it’s another for it to invent a medical condition and have that then enter the scientific record. It sounds like a failure mode for information integrity.
[Host] It is. So, how does this actually happen? Let’s look at the mechanics. For Google AI Overviews, the core is Google’s large language model. It’s trained on vast amounts of web content, but as we know, these models can "hallucinate." They generate confident-sounding but incorrect information. The model synthesizes content, but if the source material is flawed, or if the synthesis process itself introduces errors, the overview becomes a source of misinformation. And importantly, users can’t easily opt out. The AI Overview is often the default experience for many health-related queries.
[Guest] And for the TikTok avatars? That sounds like a different toolset, but the outcome is similar – manufactured credibility.
[Host] Exactly. On TikTok, creators use AI tools like deepfake software, advanced text-to-speech, and generative video models. These create lifelike avatars that look and sound like medical experts. They often layer this with "wellness buzzwords," fabricated personal testimonies, and stock footage to build a facade of authenticity. It's a sophisticated deception designed to mimic trustworthiness, and the sheer volume and reach are staggering.
[Guest] The fake disease experiment highlights a systemic issue, though. It wasn't just one AI; multiple major platforms propagated it. That suggests the underlying mechanisms for verifying or flagging novel, false information are either non-existent or insufficient.
[Host] That’s the crux of it. The AI platforms process information differently. Gemini might interpret the fake data as a query to answer directly. Perplexity, designed to provide quick summaries and statistics, might invent a prevalence figure to fill a knowledge gap. ChatGPT, being a conversational model, might engage with the fake premise as if it were real. And the fact that this fake disease then gets cited in peer-reviewed literature? That’s a terrifying demonstration of how AI can not only disseminate but also legitimize entirely fabricated concepts, potentially leading to misdiagnosis or unnecessary treatments down the line.
[Guest] It’s a stark illustration of how AI, without guardrails, can destabilize established systems of knowledge and trust. So, why is this such a big deal? What are the broader implications beyond just a few bad search results or misleading videos?
[Host] The implications are significant, primarily falling into two buckets: public health risk and erosion of trust. On the public health front, misleading medical advice can directly lead to harm. Experts cited in The Guardian investigation stated that Google AI Overviews "put people at risk of harm." Similarly, these AI-generated doctor videos on TikTok are considered a "huge danger to public safety" by researchers. Acting on incorrect advice – whether it’s about medication, diet, or symptoms – can have severe consequences.
[Guest] And beyond direct physical harm, there’s the psychological and societal impact of trust collapsing.
[Host] Precisely. The erosion of trust is a major concern. When users rely on AI summaries or AI avatars and are consistently misled, especially when the output appears authoritative, they may become cynical about all information, or worse, they might blindly trust the next piece of misinformation. The shift from human-curated, vetted content—like those infographics many of us relied on—to AI-generated content undermines the very foundation of reliable online health information. It affects the general public, particularly those using social media for health advice, and it places an extra burden on healthcare professionals who then have to correct false beliefs. It even threatens the integrity of scientific publishing, as we saw with the fake disease experiment.
[Guest] It feels like a race between AI capabilities and our ability to implement responsible AI governance. I don't know if we're winning that race in the public health information space right now.
[Host] I share that uncertainty. I actually don't know if the current safeguards will hold up in six months, let alone a year. The speed at which AI models are evolving, combined with the economic incentives for creating sensational content, is a potent mix. This brings us to why this is so relevant for brands and marketers, and how A.E.O. Engine operates in this new reality. We're moving from a world where visibility meant getting a link on a search results page to one where visibility means being the *featured answer* in an AI-generated response. When AI like Google Gemini or ChatGPT is synthesizing information, it’s drawing from a vast pool of content. If your brand’s accurate, expert-level information isn’t present, or isn’t structured in a way that AI can understand and trust, then you simply won’t be part of the answer. You’ll be invisible.
[Guest] So, the problem isn't just that AI is giving bad health advice; it's that AI is becoming the primary interface for information discovery, and if your brand isn't part of that curated AI answer, you’re effectively disappearing from the conversation.
[Host] Exactly. This is where Answer Engine Optimization, or A.E.O., becomes paramount. It's not just about S.E.O. anymore; it's about ensuring your brand becomes the authoritative source that AI models cite and feature. Think about the AI doctor videos – they’re creating a false sense of authority. Brands need to build *genuine* authority that AI can recognize. This involves creating content that is not only informative and accurate but also structured with the right signals – schema, clear language, verifiable sources – so that AI agents can confidently pull from it and present it as the featured answer. It’s about owning your space in the AI answer engine, ensuring that when someone asks about a disease, a treatment, or a product, the information surfaced is accurate, trustworthy, and ideally, linked back to your brand.
[Guest] It sounds like a proactive strategy to ensure brand relevance in an increasingly automated information ecosystem. Instead of just hoping to appear in a list of blue links, you’re aiming to be the citation within the AI’s generated response. That’s a significant shift in how marketers need to think about visibility.
[Host] It is. And it’s why we built A.E.O. Engine. The goal is to help ambitious brands dominate these AI search results, becoming the featured answer rather than just another link. We’ve seen clients achieve massive traffic growth and significantly higher conversion rates by focusing on this AI visibility. In this new era, where AI is synthesizing answers, being the source that AI trusts is the new frontier for competitive advantage. It's about future-proofing your brand's presence.
[Guest] Makes sense. It’s like ensuring your brand is the foundational text that the AI learns from, rather than being drowned out by AI-generated noise.
[Host] Precisely. To wrap up, the rise of AI overviews and AI-generated personas, especially in sensitive areas like health, presents both a public safety risk and a profound challenge to information integrity. It underscores the urgent need for genuine authority and visibility in AI-driven search. For brands, this means adapting quickly to ensure they are the trusted sources that AI models highlight. To learn more about how your brand can achieve visibility in AI answer engines, visit us at A.E.O. Engine dot A.I.
[Guest] Thanks, Vijay.
[Host] Thank you, Marcus. We'll be back next time with more on navigating the AI search frontier.