Google's AI Stumbles on Medical Advice
- Sarah Ruivivar
- 2 days ago
- 2 min read
Well, folks, it seems Google’s AI has been caught with its digital trousers down, serving up some less-than-stellar medical advice.
In a move that’s got experts raising their eyebrows and shaking their heads, Google has yanked certain AI-generated medical overviews after The Guardian’s investigation revealed some alarming blunders.
Picture this: Google’s AI advised pancreatic cancer patients to steer clear of high-fat foods, a suggestion experts deemed “really dangerous.” Turns out, this advice could actually up the risk for those battling the disease. Yikes! And if that wasn’t enough to make you double-check your search results, the AI also provided misleading info about liver function tests, potentially leaving those with serious liver conditions blissfully unaware. Talk about a health hazard!
Want to hear more? Join Mal & Matt on the Property AI Report Podcast each week!
Access from your preferred podcast provider by clicking here
In response, Google has disabled AI responses for queries like “what is the normal range for liver blood tests?” and is now in damage control mode. Spokesperson Davis Thompson assured The Verge that Google invests heavily in ensuring the quality of its AI overviews, especially for sensitive topics like health. While they claim most information is accurate and backed by credible sources, it seems a few glitches slipped through the net.
This isn’t the first time Google’s AI has been in hot water. From recommending glue on pizza (yes, really) to suggesting rock consumption, it’s been the subject of some eyebrow-raising controversies and even lawsuits.
So, what’s the takeaway? While AI is a fantastic tool, it’s a reminder that when it comes to health advice, consulting a real-life professional is always the safest bet. After all, your health is no laughing matter!
Want to hear more? Join Mal & Matt on the Property AI Report Podcast each week!
Access from your preferred podcast provider by clicking here
Made with TRUST_AI - see the Charter: https://www.modelprop.co.uk/trust-ai
Comments