Trust is the currency of healthcare. When patients trust us, they share openly, follow care plans, and stay loyal to our organizations. When trust erodes, outcomes can suffer and patient loyalty is jeopardized. Today, trust in institutions — including healthcare and hospitals — is declining. And the rapid rise of artificial intelligence is unfolding directly inside this critical moment.
AI brings extraordinary promise to healthcare. It can improve clinical outcomes, reduce friction, improve access and free clinicians to spend more time with patients. But it also introduces new risks. In the dash to adopt AI, one question that healthcare leaders should not forget is: Will our use of AI build trust, or will it accelerate its decline?
In recent months, I’ve heard comments that should give us pause. A patient told me they were considering abandoning their patient portal because they feared their data might be used to train AI. Another asked whether ambient listening in a physician’s office was capturing information for other purposes. A third expressed frustration that they couldn’t tell whether they were speaking with a human or an AI agent. Patient experience teams across the country are hearing similar concerns. Our patients are reading the same headlines about AI that we are — and they have questions.
We can debate the accuracy of these worries, but we cannot ignore the theme running through all of them: trust. If we want AI to strengthen trust rather than erode it, we must act now. Three steps are essential.
1. Establish clear, public principles for AI use.
Just as healthcare organizations articulate mission and values, they should articulate how AI will be deployed. Principles should address ethical use, transparency, stakeholder involvement and a commitment to enhancing — not replacing — the humanity of care. Leaders must embrace these principles and hold their organizations accountable for living them. When patients know what we stand for, trust grows.
2. Build governance structures that include the people most affected.
Most AI governance groups today are dominated by IT, clinical leaders and administrators. These voices are essential — but incomplete. Front-line staff and patients must be meaningfully involved in evaluating use cases, identifying risks and shaping implementation. Their perspectives can reveal blind spots that technical teams may miss. Yes, involving patients requires time and preparation. But, in my experience, the return on that investment is enormous, and few issues in modern healthcare deserve it more than AI.
3. Be transparent with patients about when and how AI is used.
Many concerns I hear would disappear with simple, proactive communication. Patients should know when ambient listening is active, when an AI agent is responding and how their data is protected. Lack of information breeds suspicion — especially in an environment saturated with AI headlines. Healthcare can be paternalistic; this is a moment to lean toward openness.
AI implementation cannot be paused — but it must be guided. It can reduce long‑standing friction, improve outcomes, and create space for the human interactions patients crave. But how we implement AI will matter as much as what we implement. If we build the right principles, governance and transparency now, we can create a healthcare system that works better for everyone — and one that is worthy of our patients’ trust.
