- OpenAI last week announced the launch of its Health in ChatGPT tool for US users, enabling them to upload their medical records to better understand their health information and track changes.
- A March poll found that 1 in 3 American adults have consulted AI regarding their health.
- 41% of those who consulted AI about health questions reported that they uploaded their personal medical information.
- OpenAI claims it will delete data from its system within 30 days to ensure user privacy.
- Some companies have made it easier for Americans to upload medical records and wearable device data to receive more personalized health information from AI.
- The legal and ethical infrastructure to handle personal and sensitive information related to AI health tools is currently lacking.
- Despite the potential benefits, experts warn that AI is not perfect and can provide problematic responses to health-related inquiries.
OpenAI's new Health in ChatGPT tool allows U.S. users to upload their medical records, aiming to enhance understanding of personal health data. With 300 million users having previously asked health-related questions, the tool seeks to centralize health information.1
However, experts express concerns about the reliability of AI in health contexts. Dr. Nick Tiller notes that AI chatbots can provide problematic responses up to 50% of the time, raising alarms about public safety. “My concern is that the public are effectively going to be the guinea pigs for this new health-specific AI chatbot,” he stated.7
Despite the potential for improved healthcare access, experts like Dr. Art Caplan emphasize the need for stronger privacy assurances and a robust legal framework to protect sensitive information. “I would never turn my data over to a private entity without more assurance of confidentiality,” Caplan warned.
The tool is not intended for diagnosing or treating diseases, yet many users may still seek health advice from it, complicating the ethical landscape surrounding AI in healthcare. As the rollout continues, the balance between innovation and safety remains a critical concern.
“As of 2023, nearly a third of Americans are considered 'medically disenfranchised,' lacking access to primary care. While 41% of users reported uploading personal medical information, experts warn that AI's reliability in health advice remains questionable, with significant risks of providing problematic responses.”
WTOP
