The legal industry’s AI landscape
AI is buzzing in the legal industry. It’s time to separate hype from reality.
I was recently invited to speak on artificial intelligence in law at the annual LMA conference along with an outstanding panel of industry experts (read my summary here). During my talk I offered a simplified view of the AI landscape of vendors and technologies in play and received great feedback and many requests for further detail.
This feedback is consistent with many conversations I’ve had regarding legal AI over the past few years: it’s hard to distinguish fact from reality, and practical examples of AI in this industry are hard to come by. Yet, we’re increasingly hearing grand claims and predictions of robot-lawyers and coin-operated legal advice dispensers at the local Walmart. And for those unfamiliar with the technical details, AI is complicated further by confusing terminology, technical jargon and a range of definitions that are sometimes applied too loosely, IMHO.
In an effort to simplify, I’ve reduced the growing list of vendors who proclaim AI super powers into a map organized by what their products actually do for the industry today:
The box sizes are not quite to scale, but do reflect how widely products in these categories are being used. And generally speaking, from bottom to top represent products that help us do existing work more efficiently to helping us make smarter and faster decisions in the work we do.
From the bottom:
Document work, such as drafting, searching, reviewing, analyzing or summarizing documents. And often times, lots of them. This is where the majority of AI applications are used in practice, primarily in e-discovery and increasingly in due diligence or large-scale contract review.
Research work, such as searching for and compiling information related to a particular argument, decision or outcome. Think beyond legal research here, as I include business-oriented research such as market or competitive intelligence gathering or client research as well.
Products in these first two categories make the work faster, better and cheaper. Literally. Less than 10 years ago, “predictive coding” and “technology assisted review” were relatively new buzzwords. But these tools and their usage is now the standard for discovery and early case assessment. There is no going back.
Looking forward, and upward, I then include predicting and automating [stuff]. OK, this is a bit vague. My point is that products in these categories often include broad capabilities that must be applied to narrow and well-defined problems and solutions. And the types of problems and solutions are growing by the day, although not all have been proven yet. So I broadened these categories for simplicity’s sake understanding that the devil is in the details.
By automating stuff I’m referring to what are often call intelligent automation or expert systems, which essentially take the knowledge of an expert in a specific area or on a specific process and capture it in software, giving the software the ability to ask and answer questions or to help with completion of tasks or projects.
And then we have predicting stuff, more often called predictive analytics or legal analytics, where given enough good data these products can identify relationships, patterns and trends and can statistically calculate a probable outcome, decision, risk or opportunity. And I emphasize good data, as this is usually the limiting factor in the success of these products.
Over the next few posts I’ll expand on these categories further, including real-world examples and my thoughts and opinions on the vendors and products in each.
Until then, please share your feedback on this model. Am I missing something, or are there other angles we should take to help us all understand the true potential of AI in the legal industry…good, or bad?
