Takeaway
AI can synthesize data and offer clinical guidance, but it can't capture the full picture of the patient in front of you. As you integrate AI into your practice, ask not only "Is this recommendation accurate?" but also "What is AI missing about this person?"
Connecting with patients | October 13, 2026 | 2 min read
By Hector Acevedo, patient, and Rahul Gorijavolu, medical student, Johns Hopkins Medicine
Imagine you’re a patient sitting across from a clinician after months of appointments, tests, and unexplained symptoms. Your medical record has become extensive, yet symptoms haven’t improved. The numbers are there. The recommendations are there. And now, AI can summarize years of encounters in seconds and suggest what might come next.
But something’s missing. The data doesn’t fully capture what those months have meant—the sleepless nights, fear of another inconclusive appointment, responsibilities waiting at home, or frustration of trying to explain that something still doesn’t feel right.
AI may know a great deal about the patient. But it doesn’t know the person. As AI becomes part of clinical care, perhaps one of the most important questions healthcare professionals can ask isn’t “Is this AI-generated guidance accurate?” but instead, “What doesn’t AI know about the person sitting in front of me?”
Correct doesn’t mean complete
AI can recognize patterns across enormous amounts of information, helping clinicians synthesize complex records, consider possibilities, and access knowledge quickly. But data and meaning aren’t the same thing. A recommendation can be technically sound and still be incomplete because lived experience, culture, family circumstances, fears, values, and barriers to care are difficult to capture in structured data.
This is where the patient’s experience outside the clinic remains essential—not simply as the recipient of a recommendation, but as a source of knowledge. When an AI-generated interpretation conflicts with what a patient is saying, what happens next? Do we become curious about the discrepancy, or does the authority of technology unintentionally make the patient’s voice easier to dismiss?
Can AI create more space for humanity?
AI promises efficiency. But, of course, efficiency shouldn’t be the central focus of exceptional care. If AI gives clinicians more time with a patient, what should that time look like? Could it become another question? More eye contact? A clearer explanation? An opportunity for a patient to say, “There’s something else you should know”? Perhaps one measure of successful clinical AI should be whether it creates more room for human connection rather than quietly replacing it. Can AI create more space for humanity?
AI disclosure
Patients should also be told when AI meaningfully informs their care. Trust becomes difficult when technology influencing a decision is invisible to the person most affected by it.
Keeping responsibility human
AI can provide information, identify patterns, and suggest possible plans for care. It can’t assume responsibility for the relationship between clinician and patient.
Before acting on AI-generated guidance, clinicians might ask:
1. Have I told my patient AI is being used in their care?
2. What does AI know about this patient—and what can it not know?
3. Does AI’s recommendation reflect the patient’s goals, values, and circumstances?
4. Have I allowed time for the patient to challenge what the technology suggests?
5. Is AI helping this patient participate more fully in their care?
6. If this recommendation is wrong, can I explain why I believed it was right for this person?