My decision to stop reporting every additional layer of comment on Judge Peck’s Da Silva Moore Opinion (see Closing down the Da Silva Moore discussion for now) was made in part because of the diminishing returns we were getting from the focus on this one case, and partly because it was becoming clouded with agendas remote from the technology and from the wider issues addressed by Judge Peck in his Opinion. There is room, however, for some straight up-and-down discussion which might otherwise be crowded out.
An example of this is an article in Metropolitan Corporate Counsel headed Can Technology-Assisted Review Coexist with Strategic Search? This is an interview with Amanda Jones, Senior Research Consultant with Xerox Litigation Services. As its heading implies, the article reminds us that the high-end search tools known variously as predictive coding, technology assisted review et al, is but one of the search tools which are available to help with the identification of the documents which matter and the relegation of those which matter less or which do not matter at all.
Xerox’s own technology assisted review tool is called CategoriX, and its broad functional description is similar to that of other players in this market (note that word “broad” – I am not saying that the products and processes are the same). In the article, Amanda Jones describes it thus:
Our CategoriX process draws upon the expertise of a limited number of attorneys and/or subject-matter experts to review and code a small sample of documents for iterative training and testing of the system. The software then extrapolates assessments across the entire review population, ranking each document according to its likely responsiveness.
A point which is easily missed by people coming new to this subject (and, perhaps, by others as well) is that keyword searching can play a significant part in, as Amanda Jones puts it, “eliminating patently off-topic documents” and by identifying documents whose importance turns on content known to be important by humans but which would not necessarily be identified as such by the algorithms. The human skill lies in ensuring that this deliberate enrichment of the starting set does not lead to a skewed or biassed test result
Whilst Amanda Jones keeps off those parts of Da Silva Moore which are proving contentious, she draws attention to the fact that keywords were used as part of the process of identifying the seed-set in that case and were used also to focus on the plaintiffs’ specific discovery requests. Keywords, Boolean searches, and concept searches have a part to play also in the QA and validation processes.
The key point, perhaps, is that this is not an either / or situation – the lawyers need to understand what each of the tools can be used for and to learn which ones to deploy at what stage in the process. This will vary from case to case. It is worth emphasising also that none of this technology simply involves throwing waves of technology tools at the problem. The human element remains vital, something which Xerox is at pains to emphasise.
Xerox were hosts to one of the best judicial panels at LegalTech, just before the specific issues and arguments of the Da Silva Moore case started to hog the limelight. The panel comprised US Magistrate Judge Frank Maas, Senior Master Steven Whitaker from the UK, US Magistrate Judge Andrew Peck, and US Magistrate Judge David Waxse.
My article about it was headed “Before they famous…” not because the participants were not well known (they all are) but because of my sense that we were about to hear very much more about predictive coding / technology-assisted review. I did not guess how much, nor at the level of detail to which the Opinion, and more specifically the Objections to the Opinion, would take us. It is, perhaps, worth catching up with that judicial discussion as it stood before Judge Peck’s opinion to ensure that the wider issues and arguments do not get lost in the case-specific comments.