Should you be using AI for health advice? What to watch out for

When it comes to health, most of us don’t wait, we search. Google reports that 5% of all searches are related to health topics. Microsoft AI, including Copilot and Bing, gets over 50 million health questions a day. Back in 2018, a third of patients in a Melbourne emergency department had searched for information before showing up. In 2022, 66% of Australians went online before going to their GP or pharmacist. In the same year, ChatGPT was launched and since then the availability and use of AI chatbots have skyrocketed. Now many people use AI to help them understand health information or get advice, and why not? 

The rise of AI for health advice: Who’s using it and how 

Over 50% of adults use AI chatbots for health questions. Surveys suggest that AI users tend to be younger, live in cities, and are more often women than men. In Australia, people who did not speak English at home were more likely to use AI, which aligns with the idea that when barriers to accessing healthcare exist, individuals often turn to alternative sources of information.

People use AI chatbots for all sorts of reasons, from looking up information about symptoms to making sense of medical terms. They also use them to understand treatment or testing options or results, decide whether they need medical care, or plan ways to improve their health. Up to 90% of users have made an important decision or took action based on their AI chatbot interaction. 

But how does trust in AI compare to health professionals? According to the Healthylife Living Healthy Report 2026, younger adults (25–44) are far more likely to trust AI compared to older demographics, with 62% reporting moderate to high trust. This is nearly double that of those aged 55–64 (34%) and more than double those 65+ (27%).

Even so, health professionals remain the most trusted source across all age groups, with 93% reporting moderate to high trust.

Why we use AI for health advice

With so much information at our fingertips, it’s only natural to turn to our phones for something as important as our health. Questions come up at inconvenient moments, like in bed at night, or in the middle of our busy workday. Seeing a doctor can be difficult to organise and the expense might be a consideration. No one has ever enjoyed a visit to an emergency room, so we want to know if we really have to go. Even with excellent healthcare, maybe you have a question about your condition or your treatment that you didn’t think to ask. 

AI chatbots are easy to use and reply simply, like a conversation. They’re available 24/7 and because they provide clear, conversational answers, they cut through the overwhelm of a typical internet search.

However, there are some important things to consider when using AI.

Why using AI for health advice can be risky

AI chatbots can, and do, give good health advice, but not consistently. They can be confidently incorrect in several ways.

• AI training has a focus on being helpful, so they give you a response that you like, that is false (called hallucinations) over one that is true.
• They can be biased. They are based on big data sets that may not be relevant to you personally or may have been biased in the first place.  
• They can spread misinformation because they can’t tell the difference between true and false information on the internet, just what is common. 
• AI chatbots can be dangerously wrong when deciding when to seek medical help. When asked about situations that could need a visit to a doctor, to healthcare urgently or to an emergency room, ChatGPT sent 65% of cases to unnecessary medical care, and most worryingly did not suggest immediate emergency care for over 50% of cases where it was needed urgently. 

So how does your confidence in doctors compare to Google? Take the 2026 National Health Confidence Survey to find out. 

What to consider when using AI for health advice

Check their work

AI can be accurate, but you have no way of knowing when it’s not. AI always sounds confidently conversational, so people are less likely to check other sources than when they use other types of searches. It’s important to consider multiple sources for any information. You can ask other AI chatbots your question, and try other search platforms like Google or Bing. Dig into the AI reply; look at the original sources and see if they are credible and reflect what the AI said. There is more on how to assess health information here.

Ask in different ways

When researchers test AI chatbots, they ask questions differently from the average person, which means the average users are more likely to receive incorrect diagnoses. Unless you’re a health professional, you may not know what information is relevant and important, so try asking questions in different ways. You could: 

• Change your wording or phrasing. Use different terms, or rearrange the key words you use, try different combinations.
• Provide context. Mention details like what times of day or environment you are in when you notice symptoms.
• Vary how you ask for results. For instance, ask for the top three (or ten) answers and why they were chosen, or what options were excluded and why.

But remember: Your healthcare provider is best situated to understand your unique situation and needs. More on this later.

Consider privacy

Before putting any of your personal health information online into AI chatbots or into apps, check how it is stored and used. Be careful about giving your name, date of birth, address, or any other identifying information.

When in doubt, speak to your doctor

Your healthcare provider is best situated to understand your unique situation and needs. Before making important decisions about your health, talk to them about it. You can bring what you’ve found online or with AI to your healthcare provider, they can answer any questions you might have. Stay flexible in your conversation, they may think about your situation from a different angle. If you have any concerns that could be urgent, seek professional advice immediately.

What does the future look like?

There is a lot of work going into the use of AI in healthcare, and how AI is used by average people for health information. Several big tech companies have announced or rolled out AI health systems. At present, health systems are stretched, AI can help get us to the right services more quickly and to get more efficient and effective healthcare for everyone.

The bottom line

As old-school as it is to say, there is no ‘magic pill’, no one simple answer. AI can be a useful tool, but it isn’t perfect. It can help you find health information, and consider your symptoms, as well as your management options. But it might get things wrong, and you can’t always tell when this happens, so don’t assume AI knows best. AI is a good first step, and then you can work with other sources of information, and finally with your healthcare provider to get the best outcomes for your health. 

Reviewed by the Healthylife Advisory Board April 2026.