The Law and Technology are not natural bedfellows. The Law explains itself in words (even words in Latin), but Technology uses networks, platforms, code – how many lawyers engaged in eDiscovery have ever seen a page of code, let alone written one? They have different mindsets, Legal and IT, and a very different experience of the world. There are even two For Dummies publications on eDiscovery: The Lawyer’s Guide to eDiscovery Technology and The Technologist’s Guide to eDiscovery Law.
Keyboard searches as a way to do eDiscovery, is that a technological or a legal solution? Neither as it turns out, because keyboard searches do not work. They are based on Boolean logic with AND, OR and NOT operators – meat and drink for IT, but hard to understand for lawyers as queries become complex very quickly, almost like programming. Even if done competently, the results are never satisfactory: you either get too few results or you end up with too many: an AND operator narrows your search, leading to fewer results, while OR broadens to get you more results – and a lot more noise too.
One of the frustrations of IT is that Legal tends to over-estimate its technological know-how. This was demonstrated by an important piece of research on the effectiveness of Boolean keyword search. Lawyers were told to stop their search when they felt confident that they had discovered 75% of all relevant documents in the data set. It turned out that, on average, they had found just 20%. But this finding did not convert lawyers to technology, or at least to better technology. On the contrary, they said: “You see, technology does not work. Manual review is still the safest option by far.”
This better technology to review documents electronically received legal sanction in 2012 when ‘computer-assisted review’ passed into U.S. jurisprudence. Since then, it has become difficult to see the woods for the trees – both with regards to the ever-changing technologies and the ever-growing terminology to describe them. Lawyers believe that words are the key to understanding Assisted Review – and being lawyers, use the confusion over terminology as a smokescreen to get the better of the opposing party in eDiscovery protocol negotiations.
IT probably does not know enough about what really matters in the Assisted Review protocol process. Here, too, the Law and Technology intersect. Take the methodology of Continuous Active Learning (or TAR 2.0) which is much less dependent than TAR 1.0 on the initial training set – a technological advance which also confers a strategic legal benefit, as it removes the need to disclose the initial training set to the opposing party before the start of the eDiscovery process.
TAR 1.0 and 2.0 are the tip of the iceberg when it comes to the jargon. It is crucial that IT as well as Legal cut through the terminology and agree what they mean when they discuss the Assisted Review process. Business and Legal can be too easily swayed by technological buzzwords such as Deep Learning and Artificial Intelligence, while some eDiscovery software vendors are complacent enough to believe that the very mention of the word ‘algorithm’ makes the case for their solution. IT’s role here is to separate what is real from what is marketing hype.