It’s not often that e-discovery vendors make national headlines, but one has this week. Without a doubt, it’s a set of headlines that Exterro, Relativity, or Nuix would be happy not to see. But HaystackID is a name now splashed across everything from Bloomberg to the NY Daily News.
The reason is, the Trump Organization has retained HaystackID to use its tools to trawl through its communications and files to meet the requirements of legal disputes running in New York. This week, Judge Arthur Engoren demanded more transparency and urgency from HaystackID, imposing every employee’s nightmare or providing regular detailed reports on exactly what they have done in the past week. In short, work that many believe should have been completed (including the Judge), seems to be moving far too slowly. In fact, at this point, it appears to be on track to miss a critical end of April deadline. Now, it may well be that HaystackID is doing the best it can under challenging circumstances. But added to this frustration is that HaystackID has past ties to the GOP (Republican Party) and indeed to the Trump Organization itself. As a result some now believe that the firm is not doing all it can to help the State of New York neutrally and impartially. Again, I personally have no reason to believe that they are doing any such thing, but it’s an illustrative instance of where tech, in this case, AI-assisted forensics and e-discovery, finds its limitations. The tech itself can analyze vast volumes of data in short order. But getting access to the sources of that data can be problematic. Furthermore, in most cases, the raw results of any analysis will be parsed, analyzed, and summarized by people before being handed over. In turn, reintroducing the potential to add flavor, restrict access and otherwise neutralize the concept of neutrality and objectivity.
In theory, at least, the technology used in the Legal Sector is ‘neutral’; it’s just a set of tools. That sense of neutrality has helped to spurn an explosive growth of technology for legal work. At the end of the spectrum, vendors and lawyers alike are touting the promise of bias-free legal analysis through the use of Machine Learning and Artificial Intelligence as core to the sector’s future. Yet, leaving questions of how truly bias-free these systems can be aside (they can never be bias-free) ultimately, we need to remember that the tools are only as good as the people using them.
More prosaically, the tech only does what you ask it to do, so you need to be careful who you get to ask the questions. It’s also possible that HaystackID regrets taking this particular job, as it’s certainly not generating the kind of publicity to make the CEO, sales and marketing, or company investors happy. And again, though HaystackID may be doing the best it can in a difficult situation, it is making itself few friends in the broader legal tech community.