I group analytics, machine learning and other advanced, systematized technologies under the umbrella of artificial intelligence. To borrow an indelible Arthur C. Clarke quote and apply it here, “Any sufficiently advanced analytics system is indistinguishable from artificial intelligence.” All the so-called AI tools we use in eDiscovery are simply advanced analytics, and generally lawyers tend to be comfortable with how analytics work. After all, email threading, concept searching, and clustering have been part of the legal industry for years, and the technology works well.
With vast amounts of data, the eDiscovery industry, particularly in Canada, has lived at the forefront of the utilization of advanced technologies. This early adoption was mostly driven by corporate clients who saw technology as a way to decrease their exposure to risk and to reduce the costs associated with managing their data; think contract management software, centralized electronic document management and so on. AI is the next step along this technological curve and is being employed in various ways within the legal industry to address, for example, contract review and analysis, dark data discovery, and document review. In general, the application of AI is designed to reduce, define and organize data.
As with any new technology unfamiliar to users, adoption can be slower than desired or needed. One of the biggest challenges around AI is getting lawyers to accept it into their workflows. Lawyers strive to meet the challenges of their clients and must feel confident in the technology they use, which oftentimes results in a preference for solutions that have been tested and considered by other members of the bar or third-party arbiters (i.e., the courts) before moving forward. This view can be limiting and detrimental to cost and outcomes.
The legal technology industry is, however, making it easier than ever not only to understand how AI tools work, but also how to best work with the tool. Workflows are now less cumbersome, and the software is significantly more user-friendly than previous versions. Software like Relativity’s Active Learning allows legal review teams to train the relevance algorithm iteratively, with the system recognizing quickly and accurately what lawyers consider relevant versus non-relevant material. A review lawyer can simply start a review queue and, after the coding of several document, the system will learn enough to begin presenting highly relevant content to the reviewer.
It’s important to note that Active Learning, like any other AI-based system, still requires human training and interaction, which specialists provide. After all, an AI system only knows the answer to a question if you give it that answer first.