Deliberate AI Deployments Require Governance and Education
During a session about building a solid foundation for AI governance and the uncertain regulatory landscape, Dr. Adam Landman, CIO and senior vice president of digital at Massachusetts General Brigham, noted that organizations don’t need to start from scratch since it’s likely they already have existing strategies that can be adapted.
“AI does introduce some new themes that need to be considered. We need to understand the validity, the accuracy. Are there bias issues? Are there equity issues? And so, we need to build upon some of the policies and governance structures,” he said.
Organizationwide education is crucial, he added, not just among leadership but also within various teams. Make sure definitions about AI, machine learning, automation, AI drift and more are effectively communicated throughout the health system and work on an understanding of what “the gold standard of AI” looks like for the organization.
Mass General Brigham has multiple layers of governance for AI, Landman said, including groups thinking about it from a high-level, industrywide perspective to the granular, tactical application of solutions. Multidisciplinary participation is key.
He also emphasized considering how different personas, such as data scientists or administrative staff members, interact with AI, and how to create safe environments in which to work with and assess the tools. Prohibiting the use of AI outright just won’t work.
DISCOVER: These AI data governance strategies will set your healthcare organization up for success.
But that also means organizations should have a strong monitoring program as AI tools continue to change.
During a separate discussion at the AI Pavilion on the expo floor, Charity Darnell, vice president and chief clinical information officer at Texas-based Cook Children’s Health Care System, offered her perspective as a nurse and a nursing leader. She said technology is often something imposed onto nurses instead of implemented in partnership with them.
So, when nurses were hesitant about the adoption of Artisight for virtual nursing, Darnell said, the health system created an innovation space where they could test the solution for themselves. The organization also chose a partner willing to work closely with the nurses.
“Give your nurses a voice and give them an opportunity to speak up,” she said.
At a later discussion at the pavilion, Jeff Sturman, senior vice president and chief digital information officer at Florida-based Memorial Healthcare System, spoke with Artisight Chief Experience Officer Dr. Stephanie Lahr about how his organization fined-tuned its vision for AI.
“We’re very specific in our goals around using AI and thinking about it from a platform standpoint,” he said.
He echoed the importance of clinician education and building governance through existing structures. For instance, at Memorial Healthcare System, he has relied on the clinical informatics governance team (consisting mostly of nurses) and the physician informatics advisory council to be peer champions and deliver updates and learnings to clinicians.
He also leveraged the data governance steering committee initially for AI, and after some restructuring and planning, the health system has closely aligned data governance and AI governance.
He added that he wants to work with partners because he doesn’t want the health system to be in the data center business, which is why it continues to move into the cloud.
Evolving Mindsets To Solve New Problems
One of the sessions during the final day of the conference emphasized once more the need for human communication as newer technologies are adopted.
“The challenge I see in health systems is having old-school thinking in the IT teams that doesn’t allow them to scale or look at a problem in a different way,” said Dr. Zafar Chaudry, senior vice president, chief digital officer, and chief AI and information officer at Seattle Children’s.
He said that the old cycle of buying new equipment to build up data centers was no longer feasible, and that technical debt was one of the major challenges of 2025.
“You cannot spend $100 million on the infrastructure. So, this is where your partnerships come in, your collaboration with other people. I think the value is now to have that cloud-based infrastructure,” he said.
