AI and Healthcare: A Budding Relationship
It’s no secret that many of us are wary of artificial intelligence (AI) taking over tasks traditionally done by humans. A recent Pew Research poll discovered that 60% of U.S. adults would feel uncomfortable if their health care provider relied on AI to diagnose and recommend treatments. But whether we like it or not, AI is already making its mark in the health and wellness sector, and some doctors are beginning to embrace its potential.
An Insightful Conversation
In this chat with Marzyeh Ghassemi, an assistant professor at MIT’s Institute for Medical Engineering and Science, and James Zou, an assistant professor of biomedical data science at Stanford University, to get the lowdown on the marriage of AI and health care. Here’s what’s happening now, what’s coming up, and what could possibly go wrong.
The Current Role of AI in Healthcare
First, let’s talk about what AI can do right now. There are over 500 medical AI algorithms and devices approved by the FDA in the U.S. These algorithms are helping clinicians make better diagnoses and assessments of patients by taking on tasks like evaluating medical images, cutting out some of the labor-intensive manual work. But it’s not just about diagnosis; Ghassemi mentioned that some models are being developed to help predict disease progression or the development of complications.
AI Empowering Patients
And let’s not forget about the patients. Medical jargon can be a nightmare to understand, so Zou and his team used ChatGPT to simplify medical consent forms, making them readable for someone at the eighth-grade reading level. Now that’s a win for patient empowerment.
Addressing the Potential Dangers of AI
But before we get too excited about AI’s potential, we need to address the elephant in the room: bias and the risk of harm. For example, one AI-driven device misdiagnosed hypoxia (oxygen deficiency) in women and minority patients, leading to undertreatment. Another algorithm used to predict health care needs for over 100 million people was biased against Black patients, as it relied on health care spending to predict future needs. With less access to care historically, Black patients often spent less, meaning they had to be much sicker to be recommended for extra care.
The Risks of Well-Intentioned AI
Even well-intentioned efforts can go awry, like when the National Eating Disorders Association had to remove its new AI chatbot, Tessa, after it provided harmful advice. As Ghassemi points out, keeping AI systems up to date and relevant with current medical advice is crucial, as it changes constantly.
Mitigating AI Risks in Healthcare
So, what can be done to reduce the risks posed by AI in health care? Tech industry leaders have called for mitigating the risk of extinction from AI to be a global priority, alongside other large-scale risks like pandemics and nuclear war. Ghassemi and Zou suggest increasing transparency, addressing bias in the data, and involving diverse stakeholders in the development process.
Moving Forward with AI in Healthcare
The future of AI in health care is undoubtedly exciting, but it’s essential to proceed with caution. We must ensure that AI serves as a helpful tool, not a harmful one, and that it benefits all patients, not just a select few. So let’s keep our eyes and minds open as we navigate this brave new world of AI-driven health care.
