Download PDF

In a past blog post, we predicted that data is arguably the products liability issue of the future, and we also explored some ways in which data can be used as a sword and a shield in products liability cases. However, arguably the most pressing, and most practically limiting issue surrounding the use of data in products liability cases is the how.

Consider for a moment just one source of data: data generated from a product’s sensors that monitor daily use, collected and stored on the user’s smartphone or computer. Smaller less sophisticated products may have fewer sensors that generate fewer data points, but there may be millions of those products in the economy. Larger, more complex pieces of machinery may be fewer in number, but they may have many different sensors and computer systems that generate a staggering amount of data from each machine.

Currently the limiting factor for plaintiffs’ counsel is the fact that this vast amount of data can be staggering. Forget what you remember about gigabytes or terabytes, this amount of data can easily be measured in petabytes. (One petabyte (PB) is equivalent to 1,000 terabytes (TB) or 1,000,000 gigabytes (GB)). When one discovery request can result in petabytes of data production, and each case may have dozens of similar discovery requests, the vast amount of data that could become relevant in a products is crippling. Sifting through this data would require expensive expert witness consultation and data management companies. For the largest and most sophisticated commercial litigation this is already commonplace, but in products cases many plaintiff’s lawyers have been cautious to wade into that level of data intensive discovery because of the vast amounts of capital that it would take to store, organize, sift, process, and present that data. This makes a case so complex and so expensive, so fast, that many plaintiffs’ lawyers pursuing products cases fall back on their more tried and true methods of telling products case narratives.

A similar, and potentially even greater financial burden is on product manufacturers who may be presented with a discovery request seeking the production of that vast amount of data. Gathering the data from many possible sources, organizing it in a format that is reviewable, and preparing it for production is a daunting task. Not to mention that the data must first be scrutinized by data experts, overseen by attorneys, to ensure that the data being produced is actually responsive to discovery. Then, there must be a review for privilege concerns. Getting the data ready for discovery responses that deal in petabytes of data is hard to even fathom in a products case. A defendant would be justified in seeking a protective order and engaging in cost shifting to the litigant seeking to compel the production of petabytes of data.

So far, the practical problems outlined above make the how a daunting task for lawyers when dealing with the production of product related data. And keep in mind, the entirety of this blog post is written around the consideration of just one source of data: data generated from a product’s sensors that monitor daily use, collected and stored on the user’s smartphone or computer. There are, of course, many sources out there.

However, it would be shortsighted to simply conclude that these practical problems mean that this is an issue that products liability litigators can ignore. With the rise of artificial intelligence, and machine automated tasks, the complex processing, organizing, and reviewing functions that drive up the cost of engaging in this kind of discovery could be mitigated. For illustration purposes, consider ChatGPT and the fact that other chat bots are reportedly now able to pass the bar exam. You don’t need a computer science degree to ask a chat bot basic questions and get somewhat intelligent responses. While that technology is still evolving, it is rapidly improving. With computer programs increasingly replacing low-level litigation tasks such as doc review, it is not unforeseeable that petabytes of data could be processed and organized at the click of a button. Clicking that button in that computer program will certainty be an expensive click, but increased computer processing capabilities will cut through some of the current barriers that exist to using data in products liability cases. And as AI improves, its ability to analyze vast amounts of data quickly will increase, as will its ability to identify patterns and insights that could be misinterpreted or missed altogether by humans.

While the practical problems such as those outlined above currently limit the use of data in products liability cases, the rapidly evolving nature of computer processing and artificial intelligence offers a path forward for making vast amounts of product data more common-place in products liability litigation. So while practical problems currently limit the use of data in products liability cases, we predict that newly evolving technologies also provide a path forward toward change.