DURHAM – Duke Health is among a leading group of health systems and payers from across the U.S. to sign a pledge advancing ethical and responsible use of Artificial Intelligence technology in health care.

The pledge, announced Thursday at the annual conference hosted by the Office of the National Coordinator for Health Information Technology, is a voluntary commitment to the principles of safety, security and trust that are fundamental to the future of AI.

“AI presents unequalled potential for advancing health with new scientific discoveries, improved diagnoses and treatment of diseases and better systems that free our workers to dedicate their expertise to patient care rather than administrative chores,” said Craig T. Albanese, M.D., chief executive officer of Duke University Health System.

“But we recognize that AI also has the potential to be misused,” Albanese said. “By signing this pledge, we are publicly stating our commitment to work toward the better good.”

Mary Klotman, M.D., executive vice president for health affairs at Duke University and dean of Duke University School of Medicine, said establishing Duke’s role as a leader in trustworthy AI has been an institutional priority for years and is foundational to advancing better health.

“This pledge actually reflects many years of work that Duke Health has already undertaken to establish the infrastructures we need to pursue AI with integrity,” Klotman said. “It puts us on record with our commitment.”

In addition to signing the pledge, Duke Health has been a founding member of the Coalition for Health AI, or CHAI, established to develop guidelines and guardrails for fair and credible applications of AI in health care.

Duke Health has also built a framework for the governance and evaluation of clinical algorithms used throughout the organization. Duke’s Algorithm-Based Clinical Decision Support framework is designed to foster innovative, safe, equitable, and high-quality patient care. This is achieved with human oversight throughout the use of an AI program to ensure that transparency, quality, and ownership are maintained.

“First and foremost, AI should serve humans,” said Michael Pencina, Ph.D., Duke Health’s chief data scientist and director of Duke AI Health. “It’s imperative that AI is developed and applied in a trustworthy manner, and we have been engaged in establishing that foundation for the last few years, as evidenced by our role in CHAI and recent publications. We are not catching up on this — it’s something that has been a differentiator in terms of being careful and proactive — and our signature on the pledge is a further confirmation of our commitments.”