AI is not your doctor: survey shows potential risks

Using google AI overviews has become part of the normal pattern of internet behaviour for millions of us. But are the overviews always accurate and shouldn’t we be concerned if vital information, such as health advice, is not correct?

Google AI overviews

It was noted last year that a new option called AI Overviews had appeared on Google search. It was also argued that Google’s new AI search offering was likely to “change the internet, for better or worse”.

Just about every provider seems to be adding AI into their products with the development and launch of OpenAI’s ChatGPT and newer competitors such as DeepSeek. There has been criticism of Meta’s WhatsApp because of its blue AI circle, which can be difficult to avoid.

It was also stated that “Google officially pushed its AI Search at its Google I/O event on 20 May, where it claimed that people using AI Overviews are finding it so useful, they’re even using search more often. In its biggest markets including the US and India, the tool is driving over 10% increase in usage of Google for the types of queries that show AI Overviews, the tech giant claimed”.

However, there have also long been fears of the accuracy of these overviews and when challenged it has been found that they are not always as good as they would appear to be…

Misleading health advice

This is particularly worrying when concerned with health information and advice, which it is of course vital to get right. A survey by the Guardian found that there is the risk of people being put at harm due to misleading and incorrect health information in Google’s artificial intelligence overviews.

The AI Overviews use generative AI to provide quick summaries of essential information about a topic or question, are said to be “helpful” and “reliable”. according to Google.

However, the survey by the Guardian found that the AI overviews often put people at risk of harm by providing false information.

There was one example which was described as “really dangerous”, by Google which saw people with pancreatic cancer being wrongly advised to to avoid high-fat foods. Experts said that doing this could increase the risk of patients dying from the disease, as it is the opposite of what should be recommended.

Google was also found to have provided false information about crucial liver function tests, which could “leave people with serious liver disease wrongly thinking they are healthy”.

In a third example, information that was described as “completely wrong” was given regarding women’s cancer tests, which it is feared could see real symptoms being ignored by sufferers.

Bad advice everywhere

This survey backs up other studies about the veracity of health information on the internet. It has been pointed out that the internet is not the same as your doctor, even though greater numbers of people are now asking AI for health advice, whether it be about wanting to know pills to take, what to eat, what treatments to try or advice about symptoms they may be suffering. It is said to feel quick, easy, and anonymous.

However, it is also judged to be a risky way of getting health information. AI tools are not the same as human doctors and don’t always know what they’re talking about. They are simply machines which get information from other online sources, some of which are reliable, while some are not.

The consequence of this is that millions of people are now getting answers to their health queries that sound clever and correct. but can actually be, as the survey in the Guardian found, really dangerous.