Artificial intelligence has moved from the research lab to the exam room, and healthcare professionals can no longer afford to watch from the sidelines. At the AIMed26 conference in Orlando, a panel of clinical leaders delivered a stark message: physicians must develop the skills to critically evaluate AI-generated recommendations, not simply accept them at face value.

The urgency stems from AI's quiet integration into routine care. X-ray interpretation, for instance, is now commonly performed by AI algorithms—often without explicit acknowledgment. Yet many clinicians lack the training to recognize when these systems produce convincing but incorrect outputs, according to speakers at the session titled "AI Is Already Treating Your Patients."

From Optional Knowledge to Clinical Competency

Dr. Orvil F. Martínez, CEO of Advance Medical Research Center and Director of the AI Coalition of Puerto Rico, emphasized that AI literacy is transitioning from a nice-to-have skill to a core clinical requirement. The challenge, he and fellow panelists noted, has shifted from developing sophisticated technology to preparing healthcare professionals and systems to use it responsibly.

Dr. Rafael J. Grossmann, a general, trauma and acute care surgeon at Portsmouth Regional Hospital in New Hampshire, stressed that successful AI adoption depends less on algorithmic sophistication than on workflow integration. Healthcare organizations must invest in comprehensive change management—including education, leadership support, clear policies, and continuous evaluation—rather than treating AI implementation as a simple software purchase.

Michael Becker, PhD, MSN, RN, Associate Principal at Press Ganey, highlighted the human factors that determine success or failure. Clinicians harbor legitimate concerns about autonomy, liability, and job security that organizations must address through inclusive implementation processes, not top-down mandates.

The Equity Gap in AI Performance

A critical concern raised at the conference involves algorithmic bias. Many AI systems were developed using datasets that inadequately represent Hispanic, Caribbean, and other underserved communities, according to Dr. Martínez. These systems may underperform precisely where healthcare disparities are already most pronounced.

Dr. Gopal K. Chopra, founding general partner at Validare Capital and a practicing surgeon, emphasized that validation in diverse populations isn't optional—it's essential for clinical credibility. AI developers must demonstrate real-world performance across varied patient populations, not just impressive metrics from controlled research environments.

Communicating AI Use to Patients

The panel addressed a delicate question: how much should patients know about AI involvement in their care? The consensus favored proportionate transparency. Simple diagnostic support requires less detailed disclosure than AI-assisted surgical procedures, but patients deserve to understand that AI plays a role and that physicians maintain oversight.

Effective patient communication should emphasize physician supervision and data protection rather than overwhelming patients with technical algorithm details. The goal is informed trust, not technical education.

The Path Forward

Healthcare leaders cannot delegate AI literacy to IT departments or early adopters. Every clinician—from residents to senior physicians, from nurses to administrators—must engage with tools already shaping patient outcomes. The alternative is a profession that has ceded critical judgment to systems it doesn't understand, serving patients who don't know the difference.

Sources

  1. Q&A: Doctors 'cannot stay on the sidelines' of AI — Healio