Google has taken away the AI Overviews feature from some health-related searches. This move comes after a report that showed that when the tool provided medical information, it could have been wrong.
The investigation prompting changes to AI Overviews
A recent study raised concerns about the accuracy of AI Overviews in answering certain health questions. For instance, searches for liver blood test reference ranges reportedly returned results without important notes on factors such as age or race. People might misinterpret their own health results because of this general information.
In response, Google has turned off these AI Overviews for the exact questions that were asked. However, searches for similar or rephrased questions still returned the AI-generated summaries at first.
The company said that an internal clinical team looked over the flagged questions and found that in many cases, the information given was not wrong and came from reliable sources. They stressed their dedication to making big changes to their systems.
Some health advocacy groups have praised the removal of these specific AI Overviews. But critics say there is a bigger problem with the system as a whole. The issue is that turning off the feature for certain search terms doesn’t answer the bigger question of whether Overviews are reliable for complicated health topics.
The main question is still whether the technology can consistently give healthcare professionals safe, detailed information. This event shows how difficult it is to use generative AI for sensitive subjects, where clarity and accuracy are very important. Google has already said that it wants to improve health-related searches by adding better overviews and specialized AI models.
