One of the difficulties in trying to persuade lawyers to adopt (or even consider) the use of technology is that they are chary of anything new. Another is that they think that their work is somehow “special”, depending on those hard-earned professional skills and not susceptible to the same solutions as everything else. It helps, perhaps, to see how analytics technology supports and enhances other industries, helping the same people to give better service at lower cost – an enabler of their skills, not a substitute for them.
I wrote recently about tracking social media for eDiscovery / eDisclosure purposes. This article expands on that theme with a look at technology developments which are specifically designed not just to track social media but to draw conclusions from it. Drawing conclusions is lawyers’ work, I hear you say. So it is, but with time and cost in mind, lawyers need all the help they can get to unearth sources which may confirm or undermine their conclusions, or point them towards things they hadn’t thought about.
Xerox is a mighty corporation whose technology interests go much wider than the eDiscovery which is my particular interest. Xerox Litigation Services is but one of the Xerox divisions, one which applies 75 years of Xerox technology research and development to a specific field. There is an obvious link from the predictive and analytical tools which Xerox develops for non-legal purposes through to CategoriX, the predictive and analytical tool which Xerox Litigation Services applies to eDiscovery. Although I use Xerox Litigation Services as the model here, the principle – that ideas developed for one industry have purposes in others – applies more widely.
The Xerox Litigation Services blog continues to provide good reading for those interested in eDiscovery. Recent articles include one about the particular implications of reviewing Mac data on a PC platform, and one on specific regulatory and compliance challenges arising from Dodd-Frank; others cover data security implications for law firms, and the use of technology assisted review. This is a site worth bookmarking for a wide range of eDiscovery topics.
The article which attracted my attention today, however, is not specifically about eDiscovery. It comes from another division of Xerox and is called Xerox tackles big data challenge: social media analytics is great, what should I do with the info? The page name (which is what caught my eye in Google), is Xerox finds true meaning in social media analytics. “True meaning” goes well beyond keywords; when lawyers read documents, they look for “true meaning” rather than particular words, though they may use the words as triggers pointing to meaning. Why not use technology to steer them towards documents whose “true meaning” points them towards the things they are interested in?
The article is about the development at Xerox Research Center of an automated data analytics platform which teaches computers to draw conclusions to evaluate the sentiment of comments beyond the mere meaning of the words. The commercial objective behind this is the ability to route customer comments quickly to the right place and with an appropriate degree of urgency. This requires an evaluation of context as well as the literal content of a tweet, blog post, Facebook entry or other source of comments.
This involves text mining, machine learning and predictive modelling which helps determine whether there is an issue or an opportunity which to be dealt with quickly and by the right person.
What has this to do with eDiscovery / eDisclosure? Well, for one thing, having documents “dealt with quickly and by the right person” is the key to efficient discovery. If the ability to include documents containing the word “project” or exclude those which include “football” is important, what about those whose language plus context indicates a conspiracy or fraud? Those planning to conspire or defraud tend not to use those words in emails – and any half-clever would-be fraudster won’t be using email anyway as newer channels – chat, Facebook, Twitter or LinkedIn – appear to offer less traceable communications. The analytical tools might, incidentally, suggest that some documents containing “project” are nothing to do with the issues and that “football” was a code word for something nefarious.
At a simple level, those same social media pots are as important a source of potentially discoverable information as Word files and emails, and the lawyer who neglects to investigate such sources may find that he or she is in breach of a duty owed to client, opponents or court; perhaps more importantly, he or she may actually have missed some evidence which matters. Given the ever-increasing volumes and the diversity of sources, it becomes important to handle all this proportionately, and any tools which can mine relevant data sources quickly and accurately can help in the battle between completeness and cost. The research being undertaken by Xerox for customer-related items has obvious application for this aspect of discovery.
It goes more widely than that, however. Lawyers seem reluctant to adopt new technology for eDiscovery purposes because, or so they say, they are nervous of delegating their duty to what they think of as a “black box”. It is helpful in that context to appreciate that the technology behind predictive coding or any other form of technology assisted review is not something new and experimental which has been developed specially for eDiscovery software. That is but one application of the technology which is in use across many industries and which is relied on as the basis for very significant commercial decisions.
The Xerox technology-assisted review tool CategoriX is also a product of a Xerox Research Center. The language of the product description is very similar to that seen in the article about the social media analytics tools – it includes expressions like “prioritise” (and its corollary “de-prioritise documents classified as non-relevant”), “issues that merit attention”, “earlier case insights and a faster path to production”. The Xerox social media analytics tools do not purport to resolve the issues or exploit the opportunities, merely to flag them and route them to appropriate people who make faster decisions as a result. Isn’t that exactly what discovery lawyers want?