LAS VEGAS — Over the course of several days at this year’s International Legal Technology Association conference (ILTACON), numerous themes and common learnings emerged as Artificial Intelligence was discussed in depth.
Some of these major takeaways included:
It’s about the data before it’s about the software
When the people who actually implement AI get together, it’s amazing how much they talk about the data itself before they get to the subject of what the software does.
A session on advanced analytics that included Karl Haraldsson of LexPredict focused more on the process of getting a firm’s data into shape and building processes to capture data (or “counting legal things,” as he put it). “Everybody wants to play in predictive analytics, but few have the data to support it,” he added. Having the right data involves professionals who understand both data and legal practice, a rare combination.
Business of Law vs. Practice of Law
Across the board, most of the examples of successful application of AI or analytics had to do with the business side of the house — not robot lawyers, but robots that help firms budget and price matters. Even representatives from IBM Watson, in their Tuesday keynote, noted that Business of Law is their focus — simply because this is what was important to the customers who have gone through Watson’s use-case identification process with them. Neil Cameron of Neil Cameron Consulting Group echoed a sentiment that is relevant here: “Before we use AI to replace what lawyers do, let’s start using it for what they won’t or can’t do” — which usually involves analyzing their business operations.
It’s only AI until you understand it — then it’s just software
This quip was repeated more than once during the course of ILTACON. Demystification of AI technology will come quickly enough as incremental successes show lawyers its practical benefits.
The Perfect is the Enemy of the Good
Another theme that emerged over and over again as speakers described their paths to AI implementation was the sentiment that “it’s not perfect.” Data can be inadequate; and analytics can give inaccurate results. This can be a hard concept for lawyers, who regularly deal with high-risk situations where perfection is the standard.
The way to get around that sentiment is to view the AI component as an input to lawyers’ work, not as the work itself. And it’s always useful to remember that human lawyers are also not always 100% perfect either.
Change in management, communication and other “soft stuff” is paramount
Jonathan Talbot of DLA Piper described his firm’s implementation of Kira for document review as “more about change management than about AI.” The collection of data on the front end, and the outputs of an AI system on the back end, all require changes to the way lawyers work, and may involve new people getting involved in the work chain.
These are not problems that the software itself solves — these are people issues that in time might even result in a fairly radical reorganization of a legal organization’s structure. Indeed, organizations need to be prepared to manage the conflicts that those shifts will entail.
Clients as resource
Several of the speakers spoke of the benefits of engaging clients in the process. Some clients engage themselves, by specifying in RFPs that they want a proposal that leverages machine learning, for example.
Law firms can proactively engage clients, however, by helping to develop data sources; by keeping pressure on firm leaders to leverage technology to increase efficiency; and by being a partner in the design of processes and workflows used by lawyers on both sides.
ILTACON’s AI sessions were well attended, with standing room only in a few. It may not be representative of the entire industry population, but at least among ILTA members there seems to be a shift in attitudes on AI, from “What is it?” to “How can I use this today?”
For more on the “unsexy” uses of AI that were discussed and demonstrated at ILTACON, see David Curle’s earlier blog post.