
Chatbots Pass Along Racist Medicine Like a Doctor Who Only Read the Oldest Textbook
Ourtake
Discredited race-based medicine, now available 24/7 through a chat window.
What happened
A Stanford-led study published in Digital Medicine found that ChatGPT, GPT-4, Google's Bard and Anthropic's Claude gave false, race-based answers to medical questions about kidney function, lung capacity and skin thickness, sometimes fabricating race-based equations. The researchers warned the tools could worsen health disparities for Black patients. OpenAI and Google said they are working to reduce bias.
Sources
- nature.com · Large language models propagate race-based medicine - npj Digital Medicine (opens in a new tab)
- nbcnews.com · Advanced AI chatbots perpetuate racist, debunked medical ideas, researchers find (opens in a new tab)
- medcitynews.com · LLMs Can Perpetuate Harmful Ideas About Race-Based Medicine, Study Shows - MedCity News (opens in a new tab)
- siliconrepublic.com · Is AI ready to be integrated into healthcare? (opens in a new tab)
- news-medical.net · Large language models perpetuate racial bias in healthcare (opens in a new tab)
- timesofindia.indiatimes.com · Study says some AI chatbots provide racist health information: Why this is dangerous - Times of India (opens in a new tab)