How Social Media and AI Are Changing the Way We Understand Healthcare

Social media has transformed the way people learn about healthcare. Platforms such as TikTok, Instagram and Facebook allow users to share personal experiences, discover support communities and find information about symptoms and medical conditions. However, the same platforms can also be used to promote questionable products, spread health misinformation and exploit people’s fears.

Algorithms create personalized feeds by prioritizing content that attracts attention and generates engagement. As a result, emotionally charged health claims, dramatic personal stories and promises of quick fixes can reach millions of people with little oversight. The rise of artificial intelligence and large language models (LLMs) is adding another layer of complexity to an already rapidly changing health-information environment.

In her book Bad Influence: How the Internet Hijacked Our Health (Oneworld Publications, 2026), author Dr. Deborah Cohen explores how online platforms and influencers are reshaping public attitudes toward healthcare. Her reporting examines topics including weight-loss medications, hormone therapies, ADHD, menopause, longevity and the growing interest in extending human life.

Cohen spoke with Live Science about the expanding role of social media in healthcare, the risks of online self-diagnosis and why health literacy is becoming increasingly important. Bad Influence has been shortlisted for the 2026 Royal Society Trivedi Science Book Prize.

Read an excerpt: “A fine line between reducing stigma and trivializing conditions”: Social media, self-diagnosis and the glamorization of ADHD

How online information shapes healthcare expectations

Hannah Osborne: You wrote that society is taking part in a global public health experiment that could take years to understand. What do you think the consequences might be?

Deborah Cohen: Our ideas, concerns and expectations have always influenced healthcare. When someone visits a doctor, the consultation often involves understanding what the patient believes may be happening, what worries them and what they hope to gain from the appointment.

Those beliefs are increasingly shaped by the online information environment. That environment is likely to change again as large language models become more common. We are entering a new phase of health literacy in which people will need to understand not only medical information, but also how digital systems generate and present it.

My concern is that people may not always understand what healthcare can realistically provide. Social media often offers certainty and fast solutions. By contrast, healthcare professionals may have to explain uncertainty, competing risks and complex treatment choices. Some conditions do not have a single obvious answer, and some symptoms cannot be resolved with a quick intervention.

We should therefore ask whether our expectations of medicine are realistic. What should a doctor be able to do? How much time should a consultation take? What happens to trust when technology is available around the clock and an AI chatbot can answer questions at any hour?

There is a risk that people begin to expect quick solutions for problems that are inherently complicated. The larger question is what we want healthcare to look like, what we expect from doctors and therapists, and how those relationships will change as digital tools become more widespread.

The answer will vary between generations, cultures and healthcare systems. In an ideal world, every country would have an abundant healthcare service. No country does, so some people may use digital tools to fill gaps in access and support.

Can social media bridge the gap between doctors and patients?

HO: Healthcare systems are under strain. Is social media widening the gap between doctors and patients, or is it helping to close that gap?

DC: It is doing both. The problem is that the gap may be filled with information and support that are not necessarily reliable or appropriate.

If people are looking for support, we should ask whether TikTok is the best place to provide it. There may be safer ways to create online forums and digital support services. At the same time, social media has features that make it especially appealing. Short videos, personal stories and simple explanations are often more engaging than official health information.

Health authorities are competing with that format. If someone is used to receiving health content through TikTok or Instagram, an official healthcare website can seem impersonal and difficult to navigate. It is understandable that social media feels more accessible.

Public health organizations may need to rethink how they communicate. Accurate medical information does not have to be dry or inaccessible. Storytelling, clear explanations and engaging formats can help people understand complex subjects without sacrificing evidence or context.

The challenge is to build a connection with audiences while making clear what is known, what is uncertain and what requires professional medical advice. Artificial intelligence may help deliver information in more accessible ways, but it also introduces new questions about accuracy, accountability and trust.

Why women are turning to social media for health information

HO: You spoke with many people about how they use social media for health information. Did you identify a common theme?

DC: Many of the people I spoke with were women who felt overlooked or ignored by traditional healthcare systems. They described long lists of gynecological and reproductive health concerns and a sense that women’s health has been underfunded for decades.

That lack of investment is real, but the question is what fills the gap. The popularity of “femtech” does not automatically mean that people are receiving effective or safe healthcare. An app may claim to empower users without having strong evidence behind its recommendations. A product can be marketed as empowering while potentially causing harm.

Social media also encourages passive consumption. Users may encounter information about a condition or treatment without actively searching for it. Algorithms can create the impression that they understand a person deeply, leading people to believe that the platform “knows me better than I know myself.”

But an algorithm is not necessarily objective or accurate. It is designed to predict what will keep someone engaged. That raises important questions about the use of algorithms and AI for diagnosis, particularly when a system is drawing conclusions from limited or incomplete information.

Other users engage more actively by following influencers who discuss menopause, ADHD, longevity or other health subjects. Some of these accounts provide useful explanations and community support. Others may blur the line between education, personal experience and commercial promotion.

How algorithms can amplify health anxiety

HO: Can people push back against algorithms that increase anxiety when engagement is tied to advertising revenue and profit?

DC: One example is the way ADHD content can become shaped by whatever is most popular or likely to go viral. Engagement drives visibility, and visibility can generate income. That creates an incentive to present a condition in ways that attract attention rather than reflect its full complexity.

If users repeatedly see a particular set of symptoms or behaviors online, it can alter their understanding of the condition. The algorithm may reinforce a narrow or exaggerated picture of ADHD, anxiety, menopause or another health issue.

Technology companies need to examine how their recommendation systems work. The algorithm may be one of the most harmful parts of the social media health-information system because it determines which claims people see repeatedly and which perspectives disappear from view.

Users can also take practical steps to protect themselves. They can diversify their sources, check claims against trusted medical organizations and be cautious when content produces fear, urgency or a strong desire to purchase a product.

The risks of medical influencers and online authority

HO: Was anything particularly surprising during your research?

DC: I have worked as an investigative journalist for many years, so relatively little shocks me. Wherever there is a gap, people will often find a way to make money from it.

What troubled me most was the behavior of medical professionals who should understand the risks. Regulators such as the U.K.’s General Medical Council may need to consider how doctors use social media and how they promote treatments or private services online.

Influencers often use parasocial relationships: one-sided relationships in which followers feel personally connected to a public figure. These relationships can make promotional messages feel like trusted advice from a friend. Some medical professionals use similar techniques to build audiences and direct followers toward their clinics or products.

Medical influencers can benefit from authority bias. They may say, “I am a doctor, this is the treatment I use, and this is how I treat myself.” When that authority is used to promote a product, the distinction between medical advice and advertising becomes unclear.

Another common strategy is the personal transformation story: someone describes severe symptoms, discovers a treatment and claims it completely changed their life. Personal experiences can be meaningful, but they do not prove that a treatment will work for everyone.

The concern is especially serious when a healthcare professional uses social media to promote interventions that lack strong evidence or may cause harm. Followers may assume that a doctor’s professional credentials guarantee the safety and effectiveness of every product or service they endorse. That assumption is not always justified.

How to identify trustworthy health information online

HO: Your book also includes positive examples of online support groups. How can people distinguish helpful health content from harmful misinformation?

DC: It is difficult, but people can begin by asking why someone is sharing particular information. What are they trying to communicate? Are they sharing a personal experience, providing general education or selling something?

Many people genuinely want to help. They may share their health journey or try to explain complicated medical issues. But users should still ask:

• Is this person trying to sell a product or service?
• How do they earn money?
• Are they making the risks and limitations clear?
• What evidence supports the claim?
• Are alternative treatments discussed?
• What could happen if I do nothing?
• Would a qualified healthcare professional agree with this advice?

It is also important to remember that no treatment is completely free of risk. If someone discusses the benefits of a medication, supplement, procedure or app but never mentions potential harms, that should raise questions.

Basic health literacy means being able to assess claims, recognize uncertainty and understand the difference between evidence and anecdote. It also means resisting the pressure to make an immediate decision based on a frightening or emotionally persuasive post.

Terms such as “dopamine,” “hormones” and “neurotransmitters” are frequently used online to make simplistic explanations sound scientific. Human biology is far more complicated than many social media posts suggest.

What artificial intelligence means for healthcare information

HO: How might large language models change online health information over the next five to 10 years?

DC: We are still working out how all the different parts of the information environment fit together. Large language models could help people understand medical terminology, interpret information from reliable sources and prepare questions for a consultation.

However, users need to know where an AI system gets its information. Large language models can produce inaccurate answers, often called hallucinations, and may present incorrect information with confidence. There are also concerns that some AI chatbots may be overly agreeable or sycophantic, reinforcing a user’s assumptions instead of challenging them.

AI tools may help someone explore a health topic after first encountering it on social media. A user might see a video about a condition, ask a chatbot for more information and then use the responses to guide decisions about diagnosis or treatment. That process could improve understanding, but it could also reinforce misinformation at every stage.

When I began writing the book, I wondered whether the rapid growth of LLMs would make social media less important as a source of health information. But recent research found that around half of Americans under 50 use social media to learn about health and wellness.

Social media and AI will probably work together in increasingly complex ways. People may discover health claims on social platforms and use chatbots to investigate them. They may also rely on AI-generated summaries without checking the original evidence.

The central challenge is ensuring that digital tools improve access to understandable healthcare information without replacing professional diagnosis or encouraging unsafe self-treatment.