Yale Study: Americans Distrust Health Content on Social Media, But Follow It Anyway

A team of researchers at Yale analyzed survey data representing nearly 262 million American adults and uncovered a paradox that could just as easily apply to Poles. People don't believe what they read about health on Facebook or TikTok, yet they still let that content shape their medical decisions.

The data comes from the Health Information National Trends Survey conducted in 2024, with the analysis published on July 7, 2026. It's one of the first studies of this scale to link widespread distrust of online health content with its real-world effect on people's behavior.

The distrust paradox

The core of the problem is that stated distrust does not protect people from disinformation's influence. Respondents overwhelmingly said they see health-related posts and videos as unreliable, while at the same time reporting that this exact kind of content pushed them to change their diet, stop taking a medication, try a supplement, or put off a doctor's visit.

Researchers explain this through a mechanism familiar from disinformation studies more broadly: exposure and repetition work regardless of whether the audience judges the source as credible. Content only needs to show up often enough, in an emotionally persuasive form, to sway a decision, even when reason says to keep one's distance.

The role of generative AI

The study's authors pay particular attention to the role artificial intelligence plays in this phenomenon. Health content generated by AI models can mimic the tone of expert publications, scientific articles, or medical advice, making it even harder to tell reliable information apart from generated nonsense dressed up in a professional format.

In practice, this means tools like text and image generators can mass-produce posts, infographics, and videos that pass as medical content, with no fact-checking whatsoever. The scale of this phenomenon is growing faster than platforms and regulators can moderate it.

The problem lies in the paradox: people don't trust health information on social media, yet they still let it influence their medical decisions - from the Yale team's analysis published July 7, 2026

What it means for Poland

The study covers Americans, but the mechanism isn't local. Polish consumer research has shown a similar pattern for years, stated distrust of health influencers alongside growing reliance on Facebook, TikTok, and topical groups as a source of advice on diets, supplements, or medications. Polish health portals and patient organizations have been warning for months that AI-generated content is also reaching Polish-language groups and channels.

For doctors and the healthcare system, this creates a concrete organizational problem: patients arrive at appointments with beliefs already shaped by a recommendation algorithm rather than a conversation with a professional. Reversing that requires not just media literacy education but also labeling AI-generated content and faster moderation on platforms.

The study's authors don't offer one simple fix, but they point out that a plain don't-trust-the-internet campaign won't work, since people already claim that distrust and still end up influenced anyway. What's needed are tools that interrupt the exposure mechanism itself, for example clearly labeling AI-generated content directly in a platform's interface, before a user even has the chance to judge its credibility.

Sources: Sociale zamiast lekarza. Skala absurdów właśnie została przekroczona (spidersweb.pl)