A Virginia health system reminds us AI does not replace the doctor: the missing piece is the evidence

By 13News Now · July 24, 2026. The original piece is a short video by local station 13News Now (Norfolk, Virginia) in which Sentara, one of the state's largest health systems, delivers a simple message: artificial intelligence cannot replace the doctor. The available material includes no figures, cited studies or direct quotes from Sentara officials, so we cannot draw out any more concrete facts than the headline offers.

That said, the message fits a trend we have been observing in the healthcare sector: medical institutions are beginning to actively communicate the limits of AI, not just its capabilities. It is no coincidence that it comes just as the adoption of AI-assisted diagnostic tools spreads across radiology, pathology and clinical triage. The more real-world hospital trials are published, the clearer the distinction becomes between what AI does well —detecting patterns in images, prioritizing cases, suggesting differential diagnoses— and what remains the doctor's domain: the relationship with the patient, clinical judgment in the face of uncertainty, and ultimate responsibility for a decision.

This institutional caution makes economic and legal sense as well as clinical sense: a health system that promoted AI as a replacement for the doctor would expose itself to a liability and patient-trust problem that no hospital wants to take on today. That is why the message that "AI does not replace the doctor" has become almost an obligatory refrain in any healthcare communication about AI, regardless of how far the underlying technology has advanced.

Our reading, with the caution demanded by a video lacking technical detail, is that the headline does not contradict the underlying optimism about medical AI: the thesis that the technology can help eradicate diseases and improve health in the long term does not depend on AI replacing the doctor, but on enhancing them. The pattern we already documented with the rollout of diagnostic AI in the United Kingdom —where real value is measured in hospitals with full waiting lists, not in laboratory benchmarks— points in the same direction: AI is gaining ground as a support tool, while medical judgment remains irreplaceable in the short term. The interesting thing will be to see, with real data from these rollouts, where that line is drawn within five or ten years.