How Is AI Revolutionizing Medicine?

When discussing artificial intelligence (AI) in healthcare, it's essential to view it as a transformative shift in the tools available to us. To give you a sense of how profound this change is, think of how algebra reshaped our understanding of mathematics and the world. In healthcare, the most significant changes are occurring in diagnostic testing, where AI allows us to interpret results more efficiently and with less expertise required. We're also witnessing a shift in how care is delivered, with traditional methods being replaced by new, AI-driven approaches.

The question of how AI will be applied in healthcare is a fascinating one. These advanced machine learning models are not the final answer to the question; they are just the beginning. AI holds immense potential for the future of healthcare. But what does that really mean? Will it take away jobs? Could it save lives? What risks does it present, and how might it affect you personally?

AI has been around in various forms for decades, but it’s only recently that people have become more familiar with it, largely due to tools like ChatGPT entering the mainstream. In healthcare, AI has been in use for years as well, primarily to help doctors transcribe notes more quickly and enhance patient interactions through chatbots. However, the rise of generative AI is truly supercharging the capabilities of healthcare professionals and scientists. AI is now aiding in predicting future health risks, making diagnoses faster and more accurate, and even assisting in discovering new drugs and treatments.

Over the past five years, AI has experienced a boom in healthcare. We've developed tools that can effectively work with vast amounts of data, allowing us to uncover new methods for treating and diagnosing diseases. Most importantly, AI is helping us tailor treatments to individual patients, predicting how specific therapies will work for each person.

Generative AI is assisting doctors in more advanced ways, but both clinicians and patients have raised concerns about the potential risks these new technologies could bring. Many people wonder how AI will change their experiences when visiting the doctor or hospital. How will it inform medical decisions? What about the risks of errors, false positives, or false negatives? Who will be held accountable for these mistakes?

These concerns are prompting a need for expertise from a wide range of professionals. To ensure responsible use, organizations like the U.S. Food and Drug Administration (FDA), the American Medical Association, and the National Academy of Medicine have been involved in establishing guidelines and codes of conduct for AI in healthcare. Currently, various groups and experts are working to make sure AI is used safely and ethically.

There's also a significant tension regarding how AI will impact clinical workflows. Clinicians play a crucial role in interacting with patients, understanding their preferences, and interpreting AI-generated outputs. Despite the efforts of major AI labs like Google and OpenAI to disrupt the healthcare industry, smaller academic labs are focusing on applying AI to specialized fields within medicine.

Since 2020, there has been a surge in venture capital investment in AI and healthcare, with around $11 billion being poured into AI healthcare startups in 2024. While these startups aim for rapid innovation, academics take a more measured approach. They focus on refining AI models and developing ethical guidelines for their use in healthcare.

Experts agree that the biggest challenge with AI lies in ensuring its responsible and consistent use. There's a need for creativity and progress, but also for clear processes to ensure that before AI is applied to patient care, it undergoes rigorous reviews to guarantee its safety, effectiveness, fairness, and equity.

Yale University has committed $150 million to AI research over the next five years. One of the prominent initiatives at Yale Medicine is the Cardiovascular Data Science Lab (or the "Cards Lab"), founded in 2020 by Dr. Rohan Care. The lab focuses on various aspects of AI, including diagnostics, precision care, and the development of therapies tailored to individual patients.

One of the lab's projects is ECGGPT, an AI tool that interprets electrocardiograms (ECGs). Traditionally, interpreting ECGs requires a cardiologist’s expertise, but ECGGPT can generate reports directly from an ECG image, making it possible for healthcare providers worldwide to get accurate readings that can be confirmed by clinicians. This AI model is able to detect conditions like left ventricular systolic dysfunction, which human readers might miss.

While advancements like those seen at the Cards Lab are progressing rapidly, there are still many hurdles before these technologies can be fully integrated into healthcare systems. Regulatory bodies may need to implement oversight and safeguards to ensure public safety before AI becomes standard practice.

Despite concerns about potential misdiagnoses, experts suggest that such instances are currently rare. Even if AI makes a mistake, a real doctor is still involved in the final diagnosis and prescription process.

So, does AI make us healthier? It's not just a million-dollar question, it's a million-lives question. AI can offer significant benefits, but it also requires wide accessibility, willingness to adopt it, and the infrastructure to support its use.

In my office, I have four photographs on the wall: the transistor, the camera, the telephone, and the source code for the worldwide web. These are the innovations in my lifetime that revolutionized healthcare delivery—think telemedicine, electronic health records, and MRI imaging. AI is poised to join this list as one of the transformative forces in healthcare.