If you’re looking to implement AI governance you need to know what a mature setup looks like, and what steps to take to get there. You need answers to practical questions such as:
What does good AI governance look like?
How can MLOps help?
What is the point of all this governance and how much is too much?
How much documentation is appropriate?
Should you have manual sign-offs?
When is an escalation needed?
What should a governance board do?
What if you are in a regulated industry?
This guide will help you to answer these questions. We do not simply state what regulators require. We will explain the trade-offs and challenges involved in AI governance so that our templates can be adapted for your organization.
Why we wrote this guide
Companies perceived as AI leaders have already landed themselves in hot water through AI governance mishaps. Unfortunately many organizations have very little AI governance in place. Even where there are data or AI governance boards, these efforts are often disconnected from data scientists and have little impact on day-to-day AI work. For governance to be effective, it must be embedded into working practices. Governance must also be pragmatic and should not unduly slow down AI projects. This guide details an approach to pragmatic governance that is embedded in the working of teams on the ground.
