As new regulations emerge across the globe, there’s increasing talk about the need for businesses to develop AI governance strategies and structures. But our research shows that 91% of businesses have no way to control the use of AI internally or in their supply chains. So how do you get started with AI governance?
Fundamentally it’s about having systems in place so you know what AI is doing in your organisation (and in your supply chain), understanding the risks and having a consistent way to deal with the many grey areas that arise when ethical decisions need to be made.
Find out what AI you have
Isn’t it obvious when you’re using AI? No – in fact far from it. You might have data science and IT people who are hand-crafting machine learning models so it could be straight-forward to ask them what they are doing and keep a record. But AI isn’t a clearly defined concept so people internally may not all define what they are doing as AI.
Plus AI is likely to be hidden inside other systems you buy in. From phone systems that use sentiment analysis to flag where call handlers have been abusing customers to HR analytics tools that recommend the “best” candidates for promotion, it’s easy to acquire a tool without management knowing it’s using AI.
Once you know where AI is in your business, just like with cyber security, you need to know where it sits in your supply chain too.
Understanding the risks
When we are brought in to work with clients, we spend a lot of time working with them to identify where the risks are and how they can be mitigated. For example, when you’re using a tool bought-in from a supplier, if the supplier’s model is built on data that isn’t similar enough to your own situation, the model could be producing misleading – or just plain wrong – results for you.
There are many ethical risks that need to be thought through and it’s not fair to leave it to the tech team to do this. It takes thinking from people with many different backgrounds to see all the angles and select the appropriate choices for your organisation. Often you’ll need to bring in people from outside the organisation who could be affected by your choices to get diverse perspectives too.
Consistent practices
If you make each ethical choice in isolation, you’re bound to be inconsistent so you need to set up a way to develop standards, make ethical choices and put them into practice in a consistent way. And the way you govern AI has to fit with the way the business works. Whilst an AI Ethics Committee could be the best solution for one organisation, it would be anathema to a very entrepreneurial company that has no other committees and where the owners are used to taking all the decisions by “gut feel”.
Transparency and training
Being open and transparent about your standards and processes is important. After all, your AI governance practices should be something you’re proud to let the world know about. And if people see them and have ideas to make them better, surely that’s all to the good?
Finally, you need to spread the word internally by training people so they know what your AI governance is all about, how it affects them and how they can influence it.
We can support you
If the prospect of sorting out your AI governance by yourself seems daunting, we can help. Just contact us for an informal discussion about your situation.
