12819
Blog
Demetri Giannikopoulos

Why You Don’t Need an AI Governance Committee

I am not, of course, advocating for a lack of governance. It is necessary for safe and responsible clinical AI implementation. What is debatable, however, is the best approach to this internal framework. 

Ask four leaders how AI is governed at their health system, and you’ll likely receive four different answers. Case in point:

  • A quarter of healthcare organizations surveyed by Deloitte in 2024 have established AI governance frameworks and dedicated teams to manage their AI models.
  • Nineteen percent of health systems surveyed by the Center for Connected Medicine in 2024 did not have standalone AI governance policies but instead integrated AI considerations into their existing frameworks.
  • Some health systems have appointed dedicated c-suite executives to oversee AI implementation, including Mayo Clinic Arizona, UC Davis Health, UC San Diego Health and UCSF Health.

For an industry that leverages best practices, this lack of consensus on an effective approach to AI governance – largely driven by a “wait and see” view of federal and state regulations – is slowing down innovation.

What Is the Best Approach To AI Governance Today?

A dedicated AI governance committee can offer deep expertise, focused attention and rapid decision-making given members’ familiarity with the technology and regulations. However, establishing and maintaining a dedicated AI governance committee can be resource-intensive and create unintentional silos.

Integrating AI into existing governance structures can streamline processes and align AI initiatives to broader organizational strategies. However, the potential for competing priorities within these committees may hinder the attention (and budget) required to address ever-evolving regulations and technology needs.

Ultimately, the best approach to AI governance is organization dependent. It should consider specifics, like facility size, technical capability, internal expertise and risk tolerance. With that said…

Here’s Why You Don’t Always Need a Dedicated AI Governance Committee

Most organizations have established clinical and operational governance processes, and these existing structures provide a solid foundation for incorporating AI considerations. Rather than creating a new workstream, focus on integration with existing processes to address the unique challenges and opportunities presented by AI.

You can do this by:

  • Identifying and addressing gaps: pinpoint areas specific to AI governance that may be lacking in your current structure, such as data transparency and bias.
  • Involving the right people at the right time: Add the internal and external expertise needed to help make decisions and build trust in the AI solution.
  • Leveraging lessons learned: Use insights from other projects to inform how AI governance may need to adapt to fit your existing structure.
  • Clearly outline roles and responsibilities: Retain a clear decision-making framework for evaluating and approving AI solutions.
  • Consider an AI subcommittee: For individuals with a passion for the subject, offer them additional room for dedicated deep dives but integrate the final decision making process into the existing structures.. 

By enhancing existing governance structures, you can effectively govern AI without creating unnecessary bureaucracy. This approach will enable your organization to stay nimble and reap the benefits of AI while mitigating risks and investing in ethical application. 

Explore More AI Governance Considerations

Governance is a core pillar of organizational AI readiness, and it’s never too early to start thinking about how it will take shape. Whether you’re just getting started with AI or have a strategy already in place, we offer a collection of resources that help foster best practice sharing. 

Explore the Latest AI Insights, Trends and Research

Demetri Giannikopoulos