By Jeff Barlow, Justice Consultant, ImageSoft
In many ways court automation—from Case Management Systems (CMS) to eFiling to workflow—have revolved around meta data. A CMS is almost entirely meta data and rules. Workflow is powered by meta data and rules. And while there is non-meta data “content” in eFiled documents, the function of eFiling systems is to use meta data to appropriately place the “content” within the documentation taxonomy.
Today, just about everyone is familiar with the term “meta data” and how it shapes, populates and drives systems. Now that we all “understand” it, meta data (and to some extent, structured data in general) is about to lose its position as the focus of information systems.
The first warning shot arrived with the advent of global search capabilities. No longer were structure, format or taxonomy required to locate specific information within a body of content. Google Search literally changed the world.
Still, even with search engines and virtually unlimited data, users had to have a pretty good idea of WHAT they were searching for. In other words, a search engine can find information for you; but generally, it can’t tell you either what to look for or what’s important. (Note: Search engines DO have some capability to judge what is most likely what you are looking for and display the choices accordingly; but that’s not the same as deciding. It is, however, quite useful for advertising.)
The new kid in town is Pattern Recognition. Human beings learn and think using Pattern Recognition.
“Quite simply, humans are amazing pattern-recognition machines. They have the ability to recognize many different types of patterns and then transform [them] into concrete, actionable steps. If you’ve ever watched a toddler learn words and concepts, you can almost see the brain neurons firing as the small child starts to recognize patterns for differentiating between objects.”
Today’s technology combines big data, deep learning and pattern recognition to enable systems to learn and “think” increasingly like humans. That in turn, presages a change in the “diet” information systems will consume. Heretofore, the staple input has been structured data arranged in a structured format. The new systems want the raw, unstructured content contained in documents, pictures, video, audio or even live visual, audio, tactile, olfactory (smell) or gustatory (taste) real time observation.
And they want a lot of it.
For court documents, two implications come readily to mind.
First, the appetite of such systems for court documents will be insatiable. An old lawyer-related parable goes like this:
Young lawyer to wise old judge: “What makes a lawyer a good lawyer?”
Judge: “Good judgment.”
Young lawyer: “How does one develop good judgment?”
Young lawyer: “What kind of experience?”
Judge: “Poor judgment.”
Pattern recognition-based systems need not just current case documents, but as much of the old documents as possible. The good, the bad and the ugly. That’s the way they learn. Like people. So, expect demand for previously un-digitized old documents, as well as old documents scanned but never converted to searchable formats.
Second, because taxonomy (the way information is organized) and usage (think workflow) continue to matter, the systems will evolve from current, rule-driven workflow to systems with even higher levels of understanding and discretion. For example, in most current eFiling systems, determination of the case class and type requires the document to carry with it some form of structured identification. Systems using pattern recognition can identify the type of document, the type of case and so on case from the body of the document.
I’m guessing that the nature of documents and document management systems will dramatically change over the next few years as pattern recognition (AKA “Artificial Intelligence”) powered systems enter the mix. And judging from history, those courts best positioned to adjust to, take advantage of and leverage the new technology will be those who have most aggressively moved to automate document management, filing and workflow with the current meta data-centric technologies available.
What do you think of this technology?