How Does Medicine Move With AI?

The future of medicine hinges on a delicate balance: leveraging artificial intelligence without losing the human touch. Dr. Nancy Oriol, a leading voice at Harvard Medical School, argued at the annual Health AI Summit on July 23, 2026, that AI should enhance healthcare while preserving compassionate, hands-on human care and personal connections. Her remarks come as healthcare systems worldwide race to integrate AI into diagnostics, treatment planning, and administrative workflows, raising urgent questions about the erosion of the patient-provider relationship. Oriol’s stance underscores a growing consensus that technology must augment—not replace—the empathy and intuition that define quality care.

The push for AI in medicine has accelerated dramatically since 2023. Tools powered by large language models now assist radiologists in detecting tumors, predict patient deterioration, and even draft clinical notes. Adoption surged after the FDA approved dozens of AI-driven medical devices, and major hospital networks, including the Mayo Clinic and Johns Hopkins, have embedded AI into electronic health records. Yet surveys show that while 70% of physicians acknowledge AI’s benefits, nearly half worry it undermines the interpersonal trust central to healing. Oriol, a former dean for community engagement, framed the issue as a matter of design: systems must be built with human connection as a core requirement, not an afterthought.

At the summit, Oriol pointed to concrete examples of AI failing when compassion was sidelined. She cited a 2025 study from the University of Toronto where an AI chatbot reduced patient follow-up adherence by 12% because it lacked the warmth of a live nurse. Conversely, programs that combined AI triage with regular human check-ins saw 90% satisfaction. Oriol stressed that algorithms can surface insights—like flagging a mental health crisis—but only a clinician’s active listening can address it. Hospitals like Boston Children’s have piloted hybrid models: AI handles scheduling and data parsing, freeing doctors for deeper conversations.

The implications reach beyond bedside manner to fundamental trust in medicine. If patients perceive care as impersonal or algorithmic, they may delay treatment or ignore advice. Dr. Eric Topol, a cardiologist and author, has called for “deep medicine” that pairs AI’s analytical power with human narrative. Regulatory bodies are taking note: the EU’s proposed AI Act includes provisions for mandatory human oversight in high-risk medical decisions, while the US Food and Drug Administration is weighing similar guidelines. Oriol’s argument aligns with a broader movement toward “human-centered AI”—a term now featured in strategic plans from the World Health Organization.

Looking ahead, the next milestones will be measured in policy and practice. By 2027, the American Medical Association plans to release specialty-specific AI integration guidelines that mandate preserving patient-clinician time. Startups focused on “empathetic AI” are attracting venture capital, with firms like Forward Health raising $100 million to develop companions that sense emotional cues. The challenge will be scaling these solutions without sacrificing the relational glue that makes medicine a calling. As Oriol concluded, ‘We must design AI that sees the person, not just the patient.’ Balanced adoption will define whether AI becomes a cold tool or a compassionate ally in the exam room.