Responsible AI governance starts where the work happens

As AI continues to permeate everyday work, leaders must decide where organisation-wide policies work and where teams need tailored safeguards. Different teams require AI for different purposes and therefore their regulation and governance should be bespoke.

A legal team working with sensitive documents and a marketing team developing campaign ideas may both use AI. However, the information they handle, the decisions they make and the consequences of an unreliable output differ.

In this video, Adam Hofmann, co-lead of Elixirr’s AI capability, explains why responsible AI governance needs to be defined for each organisation and each use case. Industry nuances play a part, but the differences within a business can be just as important.

That means designing how people use the technology, including who is trained, what they can use it for and when a person needs to have additional oversight over the result. It also means putting technical safeguards in place and evaluating outputs for accuracy, bias and consistency over time.

The aim is to create a repeatable way to make sound decisions about AI use. Leaders should be able to explain the risks of a particular use case and the safeguards chosen for it. They should also be able to identify who is accountable when those safeguards need to change.