Trustworthy AI

Chapter 11Trustworthy AI in Practice

To this point, we have investigated the qualities and nuances of the various dimensions of trustworthy AI. Armed with this insight and information, the next step is to convert knowledge into action. The discussions, strategies, and tactics for promoting trustworthy AI differ depending on the business, the function within the business, and the tool's use case. What is relevant to one may be unnecessary to another.

Recognizing that, the path forward is taking the insights from the dimensions of trust and using them as guidance for the steps the enterprise takes toward developing and deploying trustworthy AI. This is a tall order, in as much as standards, expectations, laws and regulations, and operating environments can differ between applications, business models, and geographies. Just because an AI tool is trustworthy in one instance does not mean that trust translates automatically to other use cases.

Nevertheless, despite the complexity, there are universal approaches and principles that can help an enterprise navigate this multifaceted landscape. The leading practices discussed in preceding chapters can lead organizations toward trustworthy AI, but they should be harmonized into a coherent and consistent regime that not just treats one model or use case but can govern the enterprise's entire AI strategy and operations. The methods for doing so can be grouped into three broad steps.

Step 1 – Identify the Relevant Dimensions of Trust ...