We’ve talked a lot about the perils of using generative AI, which while improving is still prone to making stuff up and exposes our data to privacy problems if used as engine fodder. But I don’t know that I’ve properly distinguished between the “bad” public technologies and the possible “good” of those installed internally – and the role information governance can play to make the latter “great.”
So let me fix that right now. Starting with:
Don’t put ANYTHING sensitive into a public engine, EVER.
Anything you feed in instantly becomes part of the reference data the engine analyzes and regurgitates, and any chance of keeping others from seeing or using it goes right out the window.
You can rewrite this story, though, by using generative AI behind your own organizational walls, for in theory, anyway, that lets you maintain control over your data so you can ensure it stays private and secure.
But wait, there’s more!
If you do even some information governance beforehand, you’ll dramatically improve the quality of the results you get OUT by improving the quality of the data you put IN. Because you’ll have vetted the source material for accuracy, timeliness, integrity, provenance, and all the other good stuff, the output has a chance to actually be great.
The old saying “Garbage In, Garbage Out” may be more applicable today than ever because we’re starting to ask our systems to repurpose the information they contain. The real trick is to keep the garbage out in the first place – which, especially in an AI context, is where information governance really shines.