Data Makes the World Go 'Round

Chapter 20Analytics and AI Governance—Managing Risk and Performance

In this chapter, we will examine the following:

• From data governance to AI governance
• Controls for AI models
• What changes with generative and agentic AI
• Laws, frameworks, and standards for AI governance
• New roles and responsibilities
• Integration with enterprise risk and compliance
• Audit considerations

There is data governance, and then there is AI governance—the two are separate but connected. In this chapter we’ll take a look at AI governance. We’ll examine why the two are different and the controls needed for AI governance as well as responsible AI. We’ll look at some of the frameworks for AI governance and some of the new roles and responsibilities needed to make sure AI governance works.

From Data Governance to AI Governance

In the previous chapter, we explored the foundations of data governance. This includes the policies, roles, and processes that ensure data is accurate, consistent, secure, and used appropriately across the enterprise. Most organizations have thought about at least some aspect of data governance, although as we saw it is a journey.

Many of those same principles are important for the governance of AI. Yet as organizations begin to operationalize AI, new forms of oversight become necessary. These include ones that extend beyond data itself to include models, algorithms, applications, and their outcomes. There are still integrity and accountability issues, but also issues of fairness ...