Across the last six months, The Health Management Academy conducted research and interviewed more than 100 health care leaders about AI governance. These leaders included C-Suite executives, informatics officers, innovation leads, and data experts across Leading Health Systems (LHS). The following insights are the net result of this research and interactions. AI Catalyst will continue to provide additional guidance and discussion opportunities on AI governance as technology advances.
Once you get past all the jargon, when Leading Health Systems (LHS) talk about AI governance, what they’re really looking for is a way to make sense of and reassert control over an unwieldy, volatile, and evolving technological trend. And that’s fair. Whether we talk about ideal committee structures, parameters for AI pilot life cycles, or methods for data bias mitigation, discussion around AI governance is grounded in a desire for clarity, actionability, and accountability.
IBM defines AI governance as the ability to direct, manage, and monitor AI activities across your organization. Simple words that underline a far-from-simple task. That’s why THMA has put together five action steps for structured, impactful AI governance
Download PDF for more information on each step.
Five action-steps for structured, impactful AI governance:
Understand what you’re trying to govern: Boost AI literacy through training, AI awareness through organizational use-case audits
Don’t reinvent the wheel: Build AI governance on your existing procurement and governance frameworks
Build awareness of AI-specific ethical challenges into training, pilot design, and evaluation processes
Don’t ignore risk mitigation strategies: Prepare to navigate worst-case AI scenarios
Orient your governance structure to rapid evolution in the AI space
Download PDF for more information on each step.