Artificial intelligence seems to be the latest must-have software for all organizations regardless of industry. Now information governance (IG) professionals are integrating artificial intelligence into their organizations’ infrastructure and IG programs.
Back when the term “big data” was coined, organizations were scrambling to build data warehouses or data lakes to try to consolidate their data and achieve a competitive advantage. But they struggled to make sense of all this data, much less to base sound business decisions on it. Now, integrating artificial intelligence can sort through these mountains of data and gain insights quickly enough to react to the ever-faster market.
So how can an IG professional help integrate AI? They can start by making the AI project team’s life easier. A 2016 CrowdFlower Big Data Scientist Survey found that typical data scientists spend 19% of their time collecting the right data sets and 60% of their time cleaning and organizing data. They do all this before designing a single algorithm. Unsurprisingly, these two activities are the least enjoyable parts of a data scientist’s day. Thankfully, these are just the types of activities an IG professional can assist with.
A mainstay of any IG program is a comprehensive data map. This data map differs from what an information technology team uses to plot out the infrastructure for an organization. An information governance data map contains information on the nature of the data and its location within the infrastructure system; a data map gives the project team a listing for all of the data sets within the infrastructure and for their locations.
The main challenge of an information governance data map is maintenance. A benefit of artificial intelligence systems is that they can be configured to assist in this task as well. These systems can regularly post reports of new data sets to prompt the humans monitoring them to update the data map as needed.
As mentioned, data scientists currently spend the bulk of their time cleaning and organizing their data. Fortunately, this task is the whole point of an IG program. Eliminating duplicates, disposing of old data and properly categorizing data sets are the program’s primary functions. This normal, everyday IG role becomes even more vital when an organization is attempting to integrate artificial intelligence.
Clean, clearly identified data sets are a major goal of any artificial intelligence program, but this goal is only half the battle. The second half lies within the algorithm driving the entire system. Algorithms are designed by humans and, unfortunately, humans can be biased toward a desired outcome. This well-known problem can be magnified when a poorly written algorithm functions at the speed with which artificial intelligence makes decisions — a business error can quickly become catastrophic.
Information governance plays a more passive but equally critical role here as well. IG is all about maintaining and distributing trusted data, so in this case, IG’s goal is to maintain a complete description of an algorithm’s function. At a minimum, the design documentation for artificial intelligence algorithms should be part of the record retention schedule, although it’s better to make it a separate document set for easier identification. The reason is that a high level of detail is required from organizations that currently fall under GDPR regulatory authority.
Information governance will be more necessary than ever in a world powered by AI. These systems are already faster than any human at performing a variety of mundane tasks. There are many ways these systems can help an organization if carefully configured from the beginning. Doing so demands the efforts of a motivated information governance team.