Where Is AI Getting Its Medical Info?

As artificial intelligence (AI) searches and conversations continue to increase and gradually take the place of traditional web search methods, it’s becoming increasingly important to understand which information these AI tools are relying on when formulating recommendations. In the context of medical searches, our clients are fighting to earn mentions through relevant, timely content that is easily digested by the AI tools. But what competition are they up against?

Dr. YouTube?

Let’s cut right to it: a study last December found that for medical-related searches, Google’s AI Overviews (Gemini) cite YouTube more than any hospital site, government resource, or medical journal. That may sound a bit shocking, but when you understand Google’s goal of keeping users on Google properties, the shock wears off pretty quickly (Google owns YouTube).

The next wave of concern might be that given the lack of peer review or any other editorial standards, YouTube is a questionable source for accurate information.

• Most people would agree that YouTube might not be the best place for people to be receiving potentially life or death information and guidance. 
• The Guardian recently ran an expose about misleading medical information within Google’s AI tools. 
• In response, Google did remove some health-related AI Overviews after that investigation, but it’s hard to tell how lasting this “fix” will be. 
• Given that over 80% of health-related searches on Google are triggering an AI Overview, that’s still plenty of opportunity for misleading information to surface and be treated as gospel.

Medical Sources in Other AI Tools

Tools like ChatGPT and DeepSeek rely far less on YouTube content as a source of information, but there are similar worries about these AI tools understanding what really makes online health information trustworthy or not. These tools pull from popular online sites such as Reddit and WebMD, but as anyone who’s spent time online should know, popularity does not equal accuracy or trustworthiness.

But people are trusting these AI tools anyway. Over half of people engaging with AI report that they trust the health information they receive, and a shocking 16% have said they ignored a doctor’s advice because AI advised them differently.

The Strategy to Earn Medical Mentions in AI

It can be super easy to bemoan how quickly the landscape is shifting and be concerned about the quality of health information people are receiving online. But the reality is the reality, and to remain part of the conversation, our clients need to post content where the AI tools are looking for it. 

In the context of AI optimization (AIO) marketing for medical practices, key tactics to become a cited source include:

• Ensure your website has updated schema, particularly for any FAQ content, that can be easily digested by AI.
• Publish new content to commonly cited platforms. RealSelf and Reddit are two places that are key to being part of the conversation.
• Whenever you post educational video content on social media, make sure that content is posted to YouTube also. In fact, a dedicated effort to generate educational content for YouTube is not a bad idea.
• Repurpose older content in new ways, for instance by answering FAQs in a video format or turning information in blogs into infographics.
• Prioritize finding and publishing “social proof” type statistics, such as how many tummy tucks you performed over the past year or your patient retention rate. This is unique content online users can’t get anywhere else.
• Continue to be part of research in peer reviewed journals and traditional local media, but make sure to capitalize on these efforts by talking them up on the platforms AI tools are most likely to cite.

A Closing Thought: What to Tell Patients

As marketers charged with bringing our aesthetic clients new business online, our focus as AI develops is how to remain part of the conversation. But there’s another important conversation to be had, and that’s when a provider and patient what AI is and is not good for. Even though it’s a tremendous tool with nearly infinite potential, the reality is that the current state of AI is simply not trustworthy enough to take the place of skilled, experienced medical oversight.

So when you start to hear from patients things like “ChatGPT said you’re good,” be happy that you’ve earned that mention, but also make sure they understand the importance of relying on the most trustworthy information, such as your own pre- and post-op instructions and guidance.