Harmonizing Data and AI Governance: To Do or Not To Do?

For over 15 years, I’ve focused on governance frameworks for data management, publishing several books and articles on the subject. Recently, professionals in the field began asking for my perspective on AI governance. To provide an informed opinion, I decided to explore the concept in-depth.

Having completed the Microsoft Professional Program for Data Science years, I was already familiar with the AI landscape. Through my investigation, I reached several key conclusions:

AI Systems are Built on Data and IT Assets

AI systems rely on data as input and output, with algorithms and enabling technologies treated as data and IT assets requiring similar governance approaches.

Limited Focus on Integrating Data and AI Governance

The integration of data and AI governance is often overlooked. It is influenced by factors like regulations, organizational culture, resources, and governance maturity.

Integration Has Advantages & Disadvantages

Harmonizing data and AI governance frameworks offers potential benefits but also brings complexities and implementation challenges.

Pros

• Unified Governance: An integrated framework provides a centralized governance structure, streamlining management processes and ensuring consistency across data and AI initiatives.
• Operational Efficiency: By reducing duplication of effort and leveraging shared standards, integration enhances efficiency in managing data and AI resources.
• Simplified Compliance: A unified governance framework simplifies compliance management by aligning data and AI initiatives with regulatory requirements.

Cons

• Lost Focus: Integrating governance frameworks can overshadow the specific needs of either data or AI governance, potentially reducing their effectiveness.
• Increased complexity: The integration process introduces additional layers of complexity, making it more challenging for teams to navigate roles and responsibilities.
• Implementation challenges: Transitioning to an integrated framework requires significant changes to processes and practices, which can disrupt workflows and demand substantial effort.

I decided to check the opinions of data management professionals by publishing two polls on LinkedIn. The

The first chart illustrates the current status of the integration: most organizations rely on either a data governance or data management framework. Relatively few have fully integrated frameworks for data and AI governance. A smaller portion has implemented both frameworks but without integration, while only a handful focus exclusively on AI governance.

The second chart reflects the vision of data management professionals on the necessity of integration. The majority believe that data and AI governance must be integrated. Some are open to integration, while a few see no need for integration or remain undecided. The charts highlight a clear gap between current practices and the desired future state, emphasizing the need for more progress toward integration.