Welcome to the Dear Laura blog series! As I’ve been working to challenge the status quo on Data Governance – I get a lot of questions about how it will “really” work. I’ll be sharing these questions and answers via this DATAVERSITY® series. In 2019, I wrote the book “Disrupting Data Governance” because I firmly believe that poor Data Governance programs are getting in the way of data programs being as successful as possible.
Read the most recent blog posts in this series here and here.
“Dear Laura,
Like the rest of the world, I geeked out over ChatGPT. Then I started doing some research and worried about how this would apply to my work in Data Governance. I’d love your perspective on this. What are the implications of AI for Data Governance for the average organization?
Thanks,
Inquisitive in Iowa”
Hey there Inquisitive,
Yes, I get the worried part too. Some days I vacillate between the excitement of finally having Rosy the robot (in case I’m aging myself, Rosy is from The Jetsons cartoon) doing my housekeeping to the terror of Ultron mucking up the planet. Don’t get me started on the impact it will have on seeking information on the internet. That’s a different topic. The key part to your question is “average organization.”
If we posit that Data Governance is accountable for usage, quality, lineage, and protection, then AI will have big potential impacts for all three (but to be fair, nuances abound). Unless your organization is “doing AI” right now (and I mean actual AI – not the stuff that sometimes gets sold as AI but is just a bunch of people doing the work behind the scenes), then the answer would be very different. For the average organization, however – for those that don’t “do AI” in any real way – the answer is quite simple: Watch and prepare.
No doubt AI will impact every organization in the world. But not tomorrow. What you can do right now to prepare is to better understand the implications on the data. AI and the machine learning that runs it need data – lots of data – and it’s not just any data. It’s good-quality data (otherwise, you’re just teaching the model the “wrong stuff”). The work of Data Governance becomes increasingly important as your organization begins to grapple with its role in AI. Where the work will come first is the protection and quality aspects of Data Governance.
There are big questions and legitimate concerns about the ethics of AI. Much has been said about ethics, and I will not embarrass myself by attempting to cover the nuances here. Seeking information is a good start. Read books like “Data Conscience: Algorithmic Siege on our Humanity” and “The AI Factor” or the venerable “Algorithms of Oppression.” Follow the authors and understand the implications, especially if you are in a field that will directly impact people or services.
The most important thing you can probably do to prepare for AI is to build a Data Quality team. Too often, I work with organizations that want to improve their Data Governance, and when I ask to talk with their data Quality team, they tell me they don’t have one. If you want a fast track to doing expensive (potentially unethical) things with AI, then sure, don’t build a Data Quality team. Otherwise, it is time for this incredibly important Data Management function to step out from the murky shadows and bask in the light.
Don’t worry: You have time to approach this exciting new technology with the right data and tools. And, if we’re lucky, Rosy will come in to clean up the rest!
Do you have a question about Data Governance you’d like me to answer? Email me at Laura at moxyanalytics dot com.