Trust Is Not a Feature of AI — It’s the Foundation

CROSS Global Research & Strategy,LLC
Aug 28
3 min read

Without trust, health care AI risks deepening divides. With it, AI could finally help close them.

Author: Dr. Shakira J. Grant

Date: August 28, 2025

Executive Summary

Trust has always been the foundation of health care, determining whether patients seek care, follow treatment, or participate in research. As AI enters the system, the same truth applies: without trust, even the most advanced tools will fail. Building trust means more than technical accuracy; it demands transparency, rigorous validation, stakeholder collaboration, and a clear commitment to augmenting, not replacing, the human workforce. Trust is not a feature of AI. It is the foundation on which its future in health care rests.

Trust and the Human Core

“If you don’t trust your doctor, you’re less likely to follow their advice. If you don’t trust the health care system, you’re less likely to enroll in a clinical trial.” Cancer research study participant

In a 2021 study of Black cancer patients, participants voiced what many communities have long felt: “we don’t trust it.” Trust wasn’t just about the research itself; it extended to physicians, teams, and the system that had historically failed them.

Now AI is entering health care with bold promises. Yet the same question lingers: have we built enough trust to ensure AI is used and used well?

The Trust Gap

Health care trust has already eroded, falling from 71.5% in 2020 to just 40 % in 2024. Recent policy changes have disrupted public health and reduced research funding, further fueling inequities in access. Layered on top are the historical harms experienced by marginalized groups within health care settings.

Adding AI to this fragile mix introduces new risks: algorithms that underperform when detecting skin cancers for darker skin tones, tools that threaten to replace jobs instead of empowering workers, and black-box systems that neither patients nor clinicians can explain.

As Stephen M. R. Covey recently argued in Forbes trust is the currency of the future, especially in the age of AI. And as STAT News reported, if medical AI tools aren’t validated through rigorous clinical trials, skepticism among both patients and clinicians will remain high.

A Black female physician with long straight black hair stands facing a patient, both connected by a glowing teal bridge of light. — A physician stands facing a patient, demonstrating the concept of trust in health care AI

Building Trust Requires Everyone

To make AI real in health care, trust must be earned at every level:

Physicians must trust tools enough to use them in practice.
Patients must trust their doctors and the AI supporting them.
The workforce must trust that AI won’t erode their roles, but instead strengthen their ability to deliver care.
Leaders must trust that listening to concerns and shaping adoption strategies with transparency pays off.

A Call to Developers: Design for Trust

Developers sit at the frontlines of shaping whether trust is built or broken. Four principles stand out:

Design for Transparency, Not Just Accuracy. Explainability matters. Even partial insights into how a system reached its decision can strengthen confidence.
Validate With Rigor. Benchmarks aren’t enough. Clinical-grade trust requires rigorous trials and supporting data that mirror the standards expected of bringing a new drug or device to market.
Co-Create With Stakeholders. Build with physicians, patients, and health care workers, not just for them. Trust grows when people see their feedback reflected in the tools they use.
Augment, Don’t Replace. AI that reduces burdens and empowers clinicians builds trust. AI that threatens jobs undermines it.

Trust Is Not Optional

AI is not a special case, it’s the latest test of a health system already in a trust crisis. Trust has always been a determining factor in whether people engage in care, follow treatment, or participate in research studies. Without it, AI’s benefits will bypass those already on the margins—racial and ethnic minorities, people with disabilities, and others historically excluded.

The call to action is clear: trust is not an add-on feature. It is the foundation. Without it, AI will deepen divides. With it, AI might finally help close them.

Want to stay on top of the latest insights in health care AI? Join our community to receive updates on our blog series and other resources.

Subscribe Here