Medical Schools Urged to Embrace AI in Training Future Doctors

Patients increasingly consult artificial intelligence tools like ChatGPT before their medical appointments, prompting a call for medical schools to integrate AI training into their curricula. Angelo Volandes, a professor at Dartmouth’s Geisel School of Medicine, emphasizes that the healthcare system is preparing future doctors for a world that no longer exists.

Despite the growing reliance on AI in healthcare, many medical students and residents are prohibited from using these tools in their coursework. This restriction raises concerns about the readiness of new physicians to engage with patients who arrive informed by AI-generated information.

AI’s Role in Modern Healthcare

Some institutions are leading the way in adapting AI education. For instance, Harvard Medical School has initiated a Ph.D. track in AI Medicine, while Dartmouth is embedding AI literacy into clinical training from the outset. Yet, the overall pace of change is insufficient given the rapid expansion of medical knowledge.

Every day, hundreds of new medical studies emerge, particularly in fields like oncology, making it increasingly challenging for individual clinicians to stay current. “Within a decade, clinicians who do not use validated AI tools will struggle to defend their treatment decisions in malpractice cases,” Volandes warns. He notes that patients often arrive with questions about treatment options that their physicians have not considered, highlighting a significant gap in knowledge that AI could help bridge.

A recent example illustrates this shift. A patient in Boston consulted a chatbot that suggested three treatment options the physician had not initially considered. Together, they explored these alternatives, which ultimately contributed to a positive outcome for the patient. While AI offered information, it was the doctor who provided the necessary emotional support, demonstrating the complementary roles of technology and human care in medicine.

Proposed Changes to Medical Education

Volandes argues that medical education must evolve to prepare students for this new reality. He proposes the introduction of several key components to enhance AI integration:

1. **AI Verification Protocols**: Medical schools should implement rounds where students present AI consultations, detailing the model used, recommendations made, and decisions overridden. This practice would be documented and reviewed by attending physicians, fostering accountability and transparency.

2. **Transparency Standards**: The Accreditation Council for Graduate Medical Education (ACGME) should mandate that residents document AI consultations similarly to specialist referrals. This includes recording the questions asked and the clinical judgments made based on AI feedback.

3. **Competency Assessments**: Medical licensing boards ought to incorporate AI literacy into their examinations. Future physicians should understand which AI models are validated for specific clinical questions and be aware of their limitations.

4. **Patient Consent Frameworks**: Patients deserve to know when AI influences clinical decisions. Volandes emphasizes the importance of training students to communicate effectively about AI’s role in their care, ensuring that patients are informed partners in the decision-making process.

Volandes highlights the unique position of Dartmouth Health, which serves rural communities in New Hampshire, Vermont, and Maine. With a severe shortage of specialists in critical areas, such as geriatrics and palliative care, the integration of AI into medical training can help bridge gaps in care.

As medical students and residents prepare for interviews at training programs, Volandes urges them to inquire about AI training and its application in clinical settings. They should demand education that equips them to answer patients’ AI-generated questions and understand the tools that can enhance their practice.

Volandes calls on colleagues in academic medicine to advocate for the establishment of AI competency standards by 2026 and to replace existing bans on AI with constructive protocols. He insists that students deserve training relevant to the evolving landscape of healthcare, rather than outdated methodologies.

The integration of AI into medical practice is not about replacing human doctors; it is about equipping them with additional tools to improve patient care. The future demands that medical schools and health systems adapt quickly to ensure that physicians are ready to meet the needs of their patients in an increasingly technological world.