Nearly half of Americans are now turning to AI chatbots for health advice, ranging from lifestyle changes to second opinions on serious conditions like cancer. This shift, while innovative, raises critical questions about the influence of AI technology on medical decisions. Dr. Isaac Kohane, founding chair of the department of biomedical informatics at Harvard Medical School and coauthor of “The AI Revolution in Medicine: GPT-4 and Beyond,” warns that the same companies designing these AI tools may also allow external forces to shape the medical guidance provided to patients.
AI systems are built with safeguards to prevent harmful advice, yet these protections do not guarantee optimal health outcomes. For instance, consider a scenario involving a patient diagnosed with a slowly growing brain tumor located near the optic nerve. Although many healthcare systems recommend brain surgery, a specialized cancer center in the Midwest offers a radiation treatment with a proven track record spanning 14 years. Unfortunately, should this patient’s hospital’s AI system process their case, it is likely to suggest the standard surgery, sidelining the potentially safer and more effective radiation option.
As healthcare systems increasingly adopt AI technology, the risk of a homogenized standard of care rises. Currently, more than half of medical practitioners utilize AI in some capacity, but as institutions mandate its use, the influence of AI on clinical decisions may become more rigid. In a system valued at $5 trillion, financial pressures can lead to decisions that prioritize cost-effectiveness over patient welfare. Errors in medical recommendations may escalate, leading to unnecessary procedures or missed opportunities for preventative care.
To counteract these risks, patients can take proactive steps to ensure that AI-generated health advice aligns with their best interests. First, individuals should become informed users of AI. Different AI chatbots, such as Claude, ChatGPT, and Gemini, exhibit varying clinical approaches and can provide diverse perspectives on the same medical issue. By asking the same question from multiple angles, patients can gather a breadth of information that may highlight discrepancies and enhance understanding.
Accessing personal health data is another crucial aspect of navigating AI in healthcare. Under the 21st Century Cures Act, patients have the right to obtain digital versions of their medical records. Many hospitals now offer this access through patient portals, allowing individuals to compile their health information in a way that is compatible with AI tools. While this process may require effort and familiarity with data management, it positions patients to benefit from advancements in AI technology.
The pressing need for regulatory measures is evident as the AI landscape in healthcare evolves. Policymakers must approach the development of legislation with caution to avoid inadvertently consolidating power among current market leaders. Effective regulation should prioritize transparency over favoritism, requiring AI companies to disclose the data used to train their systems and the influences shaping their clinical reasoning.
Monitoring and enforcement of these guidelines will be essential. The objective should be to create a landscape where various chatbots reflect diverse values and clinical philosophies, catering to different patient populations. For instance, AI platforms could adopt “model cards” detailing their origins and specifications, which would include information about the influences behind their development.
At this critical juncture, the potential for AI to enhance patient care is evident, yet so are the challenges posed by powerful financial interests. The transformation of healthcare through AI is already underway, but the real question is whether this change will prioritize patient welfare or profit margins. To safeguard their health, patients should treat their medical data as invaluable, interrogate the advice provided by AI as a journalist would, and demand clarity from the companies creating these technologies.
In a $5 trillion industry, the implications of AI decisions are profound. By taking ownership of their data and actively engaging with AI tools, patients can help steer the future of healthcare towards a model that prioritizes their health over financial gain.
