OpenAI has launched ChatGPT Health, a specialized feature designed to cater to health and wellness inquiries. This initiative responds to the increasing demand for AI assistance in health-related questions, with an estimated 40 million daily queries directed to ChatGPT. In an interview with Northwestern Now, Dr. David Liebovitz, co-director of the Institute for Artificial Intelligence in Medicine’s Center for Medical Education in Data Science and Digital Health at Northwestern University Feinberg School of Medicine, outlined the potential benefits and challenges associated with this new tool.
Understanding how patients interact with AI is crucial in today’s healthcare landscape. Dr. Liebovitz emphasized that the main consideration is not if patients will utilize AI for health information, but how to ensure they can do so effectively and safely. He pointed out the necessity for appropriate guidelines and realistic expectations regarding AI capabilities.
ChatGPT Health represents a significant opportunity for patients to access and interpret their medical records more effectively. The 21st Century Cures Act mandates that healthcare systems provide patients with complete access to their medical records through standardized application programming interfaces (APIs). According to Dr. Liebovitz, tools like ChatGPT Health can assist patients in comprehending their data without incurring any additional costs. Patients can understand lab results, prepare questions for healthcare appointments, and identify potential gaps in care.
Dr. Liebovitz described a vision where patients could download their medical records using APIs provided by healthcare organizations and run them through AI models on personal devices. This would enable personalized insights without involving third-party servers, eliminating concerns about subscription fees and privacy trade-offs.
The potential of AI tools extends beyond mere patient engagement. Dr. Liebovitz highlighted the long-standing issue of diagnostic errors and care gaps in medicine, which were documented over 25 years ago in the Institute of Medicine report, “To Err is Human: Building a Safer Health System.” AI assistants that can review a patient’s comprehensive history and flag concerns offer a significant advancement compared to traditional methods that rely on patients’ internet searches.
Despite the promising potential, Dr. Liebovitz cautioned about the risks associated with using ChatGPT Health. Health data shared with this AI tool is not protected by the Health Insurance Portability and Accountability Act (HIPAA). Unlike communications with healthcare professionals, which are legally privileged, interactions with AI tools may be subject to legal scrutiny in litigation. This raises critical concerns, particularly for sensitive health matters such as reproductive or mental health.
Addressing patient privacy is essential in the development of AI tools. Dr. Liebovitz proposed an alternative approach that mitigates privacy concerns by enabling AI models to run locally on a patient’s device. Modern smartphones possess sufficient processing power to operate advanced language models without transmitting data to external servers. This method ensures that patients’ health information remains private and secure, avoiding risks associated with cloud storage and corporate data handling.
The research group at Northwestern University is actively exploring ways to make these privacy-conscious solutions a reality. With access to standardized health records, powerful mobile hardware, and advancing open-source models, the focus is on empowering individuals to obtain valuable insights into their health data while maintaining control over their information.
In conclusion, as OpenAI’s ChatGPT Health takes its place in the health technology landscape, the intersection of AI and patient care continues to evolve. Striking a balance between innovation and privacy will be crucial as healthcare systems adapt to these new tools, potentially transforming how patients engage with their health information.
