Thanks to the web, there’s never been more health information at our fingertips, which can be a blessing and a curse. Noticed a strange new ache? You can head straight to Dr. Google — or to TikTok, which can introduce you to a condition you’ve never heard of. And now, AI tools like ChatGPT can answer complicated health questions almost instantly.
In theory, having greater access to wellness info within seconds should make us better informed, right? But in practice, it can make figuring out what to believe surprisingly difficult.
“We're living in an explosive digital age,” says Aasma Shaukat, MD, a gastroenterologist at NYU Langone Health. Traditionally, health literacy meant being able to understand and use information from your medical provider — like deciphering a prescription label, preparing for a test, or following your doctor’s instructions. Today, Dr. Shaukat says, health literacy increasingly requires media literacy, too: “Who is making a claim? What evidence supports it? Does the person sharing it have a financial incentive? And, crucially, does the information actually apply to you?”
Here’s how to become a savvier consumer of health information — whether it comes from a headline, social media post, study, or AI chatbot.
Start by asking who's behind the information
Before focusing on what a piece of health content says, consider who’s saying it.
Dr. Shaukat recommends looking for clear sourcing, appropriate expertise, disclosures of potential conflicts of interest, and references to established guidelines or public health institutions. You should also consider whether the source presents both potential benefits and harms rather than making an overwhelmingly one-sided case.
Certain language should immediately make you skeptical. “Be wary of claims involving miracles, breakthroughs, 'no downsides,’ or absolute certainty,” Dr. Shaukat says. Personal testimonials can be compelling, but someone saying “this worked for me” isn’t scientific evidence. And conspiracy-minded phrases like “your doctor doesn’t want you to know this” should be another major red flag.
The same need for scrutiny applies even when research really does exist. If a company funded a study of its own product, for example, that doesn’t automatically make the results invalid. But it does make it important to understand how the study was designed and whether researchers took steps to reduce potential bias.
Don’t stop at the headline
Even a technically accurate headline can leave out important context.
Dr. Shaukat points to colorectal cancer screening as an example. A headline might announce that a new blood test can “replace” a colonoscopy. But that can give readers a very different impression from what the evidence and guidelines actually say. As she explains, “FDA-approved blood tests exist for colorectal cancer screening, but other screening methods remain preferred options.” According to the American Cancer Society guideline, blood-based tests should be recommended only to individuals who decline or do not complete preferred screening tests, like colonoscopy or stool-based tests. And regardless of what non-invasive screening method a patient may choose, a positive result always requires a follow-up colonoscopy.
It’s important to remember that screening options can’t simply be categorized as “good” or “bad.” You'll want to talk with your healthcare provider about how well it detects disease, its false-positive and false-negative rates, whether it can detect abnormalities before they become cancer, what happens after a positive result, how often testing needs to be repeated, and whether the option makes sense for your individual risk.
In other words, the question isn’t only, Does this test work? It’s also, How well does it work, compared with what, and is it right for me?
Taking an oversimplified claim at face value can have real consequences. Dr. Shaukat says misinformation about screening could lead someone to choose an inappropriate test or delay care that would be better suited to their individual risk.
Understanding FDA approval
Many people take the phrase “FDA approved” as a strong signal that they can trust a product. However, not everyone may be familiar with what exactly this important regulatory label means.
“The FDA evaluates whether a product is safe and effective for a specific intended use and population,” Dr. Shaukat explains. It's important to understand that approval doesn’t necessarily mean a test or treatment is the best choice for every patient, that it improves long-term outcomes at a population level, or that medical guidelines recommend it as the best option available.
That’s why Dr. Shaukat suggests that in addition to looking for FDA approval, people check what professional guidelines say. Guideline-making organizations like the USPSTF (U.S. Preventive Services Task Force) or American Cancer Society weigh the available evidence before determining whether something should be recommended, considered in certain circumstances, or avoided.
What about asking AI?
AI adds a complicated new dimension to health literacy.
Dr. Shaukat says tools like ChatGPT can be genuinely useful as a starting point. If you’ve recently been diagnosed with a condition you know nothing about, AI can help you learn about causes, risk factors, lifestyle changes, medications, and potential side effects. It can also help explain unfamiliar terminology or generate questions to bring to your next appointment.
Someone might even enter their symptoms and ask for possible explanations, she says. It helps to give useful context such as your age, when symptoms began, and what makes them better or worse — provided you are careful not to share any sensitive or personally identifiable information with public AI tools.
But there’s an important distinction: AI can generate possibilities; it can’t replace a doctor's diagnosis or exercise clinical judgment.
“AI systems rely on the information they’ve been trained on, and users may not know exactly what that includes,” explains Dr. Shaukat. They also can’t fully account for an individual’s medical history, risk factors, physical exam results, or extenuating circumstances in the way a clinician can. So even if an AI response sounds authoritative, that doesn’t necessarily make it accurate.
To check an AI response, ask it to provide links to sources — then read those to see whether they actually support the chatbot's answer. You can also specifically request information based only on scientific or medical guidelines and peer-reviewed research published within the last five years. (Though that search doesn't guarantee accuracy on its own.)
Most importantly, treat the answer as the beginning of your research, not the end of it. If an AI tool tells you something that conflicts with what your doctor has told you, Dr. Shaukat’s advice is straightforward: Talk to your doctor and defer to their medical guidance.
Not everyone has equal access to good information
Of course, knowing how to evaluate health information assumes you can find trustworthy, accessible info in the first place.
In research examining colorectal cancer information on social media for people whose primary language isn’t English, Dr. Shaukat and her colleagues found an "alarmingly low” number of reliable sources. There were few gastroenterologists producing content in non-English languages and limited balanced information from other organizations.
Even when credible information existed in Spanish, much of it was written rather than presented in video — an important distinction because being able to speak a language doesn’t necessarily mean someone has the same level of reading fluency. Resources in other languages were even more limited.
Interestingly, this may be an area where AI could eventually help. Dr. Shaukat envisions tools such as AI-powered navigators that could communicate health information in a patient’s preferred language. But the same rule applies: The technology is only useful if the information it delivers is reliable, balanced, and based on good evidence.
That’s a good standard to remember when looking for almost any health data online. You don’t have to ignore social media, skip health headlines, or swear off AI. Instead, ask where the information came from, look for evidence and context, question claims that sound too certain, and use what you learn as a jumping-off point for a conversation with a healthcare professional. Because ultimately, the goal isn’t to become your own doctor: It’s to become a better-informed patient.
Dr. Shaukat is a consultant for Freenome, UniversalDx, and Motus GI.
Abbott offers both stool- and blood-based colorectal cancer screening tests. Talk to your doctor to understand your eligibility for screening and determine what solution is right for you.