AI in Medicine: Could Technology Outperform Human Doctors?

Advancements in artificial intelligence (AI) are leading to discussions about whether technology could surpass human doctors in diagnosing and treating patients. According to Charlotte Blease, an associate professor at Uppsala University and a researcher in the Digital Psychiatry Program at Harvard Medical School, the limitations of human practitioners may open the door for AI to play a more prominent role in healthcare.

The medical field is currently grappling with alarming statistics. Misdiagnoses are a significant issue, with medical errors accounting for more than a quarter of a million deaths annually in the United States alone. Each week, the equivalent of four airplanes, each with 170 passengers, crashes without making headlines. Blease emphasizes that diagnostic errors affect most people at some point in their lives, with approximately 22 million patients in Europe suffering from undiagnosed rare diseases. The situation is exacerbated in low- and middle-income countries, where misdiagnosis rates are expected to be even higher.

It is not just diagnostic errors that plague the healthcare system. Despite modern medicine’s scientific foundations, evidence-based treatments are provided only about half the time. Moreover, those who require medical attention the most—such as the elderly, low-income individuals, and marginalized communities—often face significant barriers to accessing care. The American Time Use Survey indicates that patients typically spend an average of two hours for a 20-minute visit, with low-income individuals facing even longer wait times.

Blease notes that doctors are facing unprecedented challenges. Burnout is rampant, with half of all U.S. physicians reporting feelings of exhaustion and 20% suffering from depression. Tragically, between 300 and 400 doctors in the U.S. take their own lives each year, a sobering statistic that equates to the loss of an entire graduating medical school class annually.

The United Nations projects that by 2037, the global population will increase by a billion people, raising further demands on an already strained healthcare system. By 2050, most individuals aged 50 or older in the U.S. are expected to manage one or more chronic illnesses. Compounding these issues is the rapid pace of medical knowledge advancement, which takes an estimated 17 years to translate from research to clinical practice. With a new biomedical article published every 39 seconds, it is virtually impossible for doctors to keep current with the latest developments.

Given these challenges, AI presents a compelling alternative. Unlike human practitioners, AI does not suffer from fatigue or distraction and can process vast amounts of data continuously. Preliminary studies show that AI tools can outperform human doctors in clinical reasoning, particularly in complex cases. For instance, an AI chatbot recently identified 90% of rare disease diagnoses correctly when presented with clinical cases, significantly surpassing the performance of human doctors.

While AI is not without its own biases—reflecting societal prejudices embedded in its training data—Blease argues that addressing these biases in AI may be more feasible than eliminating the inherent biases present in human decision-making. A study conducted in 2024 revealed that AI models were more likely to diagnose men than women for common conditions and recommended fewer scans for Black patients compared to white patients. However, research also shows that AI can identify discriminatory language in medical records, offering a potential avenue for improving equity in healthcare.

Patients may also find themselves more comfortable discussing sensitive issues with AI than with human doctors. Historical instances, such as the first recorded patient interaction with a computer in 1966, indicate that individuals often feel freer to disclose personal health information to machines. This tendency could enhance patient engagement and lead to more accurate assessments of their conditions.

As the conversation around AI in healthcare continues to evolve, Blease’s insights in her new book, Dr. Bot: Why Doctors Can Fail Us—and How AI Could Save Lives, highlight the potential for technology to bridge gaps in the current medical system. While the thought of entrusting healthcare to AI may seem unsettling, the possibilities for improved accuracy, efficiency, and equitable treatment warrant serious consideration. The future of medicine may very well depend on a harmonious integration of human expertise and technological advancement.