Let’s start defining “governance” as a set of rules, processes, frameworks, and tools within an organization to ensure that what we are doing aligns with internal principles and values, legal requirements, and social and ethical standards.
When it comes to AI we are talking about everything we need to evaluate the right development, use and adoption are in place.
As well as other IT/Digital systems, also AI solutions have a lifecycle, from “Planning & Design” to “Operation & Continuous Monitoring”. AI Governance needs to cover all the lifecycle’s steps.
To put in place a good AI Governance we need:
- Right stakeholders (AI Team, Dev Team, AI Lead, Legal Team, Customer and Coworker Success Team)
- AI Policy (rules, values and guidelines)
- Risk categories (in EU based on EU Artifical Intelligence Act)
- AI Registry (to maintain always updated)
- Solid AI Literacy plan (general mandatory training for any coworker)
- AI Adoption guidelines (specific training plan, clear business value, KPIs or OKRs, etc, on a single AI project basis)
Luckily we don’t need to reinvent the wheel. There are several AI Governance frameworks already available and ready to use.
There also many other frameworks available from IT vendors, maybe they are more actionable, but strictly connected to owned specific solutions.