Google removes AI-generated summaries for select medical queries.

Earlier this month, an investigation revealed that Google was delivering misleading and even false information through its AI-generated responses to certain medical queries. Following this, those problematic results seem to have been taken down. The original findings highlighted some concerning examples:

> In one instance deemed “really dangerous,” Google incorrectly advised individuals with pancreatic cancer to steer clear of high-fat foods. Experts pointed out that this recommendation was completely wrong and could potentially heighten the risk of mortality for patients battling the disease.

> In another troubling case, the AI provided inaccurate information regarding essential liver function tests, which could mislead individuals with serious liver conditions into believing they are in good health.

As of this morning, the AI responses for queries like “what is the normal range for liver blood tests?” have been completely disabled. Google has not commented specifically on the removal. A spokesperson mentioned that the company invests heavily in the quality of AI responses, especially in health-related topics, asserting that most of the information is accurate. They noted that their internal team of clinicians reviewed the shared information and found it to be mostly correct and supported by reputable sources. When the AI responses lack context, they aim to make broad improvements and take appropriate action under their policies.

This incident adds to the ongoing controversies surrounding a feature that has previously suggested bizarre advice, such as putting glue on pizza and eating rocks, and has faced multiple lawsuits.