Machine Learning and the Legal Industry

In 1970, a Japanese roboticist named Masahiro Mori coined the phrase “the uncanny valley.” Even if you’ve never heard the phrase before, you’ve likely experienced the phenomenon.

The uncanny valley describes the unease we feel when we look at something that’s almost human, but not quite. Mori observed that as robots get closer to looking like people, the human emotional response becomes increasingly anxious. When something almost looks right but still seems “off,” it makes us uncomfortable.

That same sense of unease can extend beyond the visible. For many, the idea of Artificial Intelligence (AI) is fundamentally unsettling. Maybe it’s all the sci-fi movies we’ve seen that suggest computers will take over the world. Or perhaps it’s an inherent distrust of machines. They aren’t perfect, after all.

Whatever the case, an early form of AI is starting to gain some real traction. Machine learning has the potential to dramatically change multiple industries. In our last post, we explored how machine learning is reshaping the medical field. This time around, we’ll examine how machine learning could impact the legal industry.

Automated Research Assistance

Research is one of the core tasks for any law firm. Every case includes an extensive period of research that takes into account previous cases, relevant laws and legal precedents. Attorneys spend tremendous amounts of time pouring over huge amounts of information. Anything that can accurately speed up that process would be beneficial.

That’s where machine learning comes into the picture.

As Innovation Enterprise observed in a recent article, machine learning is well suited for processing complex data, capable of identifying patterns and even inferring rules. The technology has already proven this capacity in other industries. It only makes sense to apply the same kind of functionality to the legal industry. This would give attorneys a budget-friendly way to overcome a major slow-down in their process.

Boosting Efficiency

There is, however, a caveat to the idea of automating legal research with machine learning. The technology looks promising, but not perfect. By that we mean we’re a long way off from being able to turn legal research completely over to machine learning.

In this sense, machine learning can be thought of as a significant boost to the efficiency of the research process, but not a solution or replacement for human interaction. Machine learning can do a preliminary sweep of potentially relevant data. However, an attorney will still need to make final strategic decisions about what relates to the case at hand and what does not.

Avoiding Overfitting

A possible pitfall of machine learning as it applies to research is what statisticians call “overfitting.”

In his paper for The Washington Law Review, Associate Professor of Law Harry Surden said this about overfitting: “The general idea is that it is undesirable for a machine learning algorithm to detect patterns in the [data] that are so finely tuned to the idiosyncrasies or biases . . . that they are not predictive of future, novel scenarios.”

In other words, machine learning is limited to the mathematics of algorithms. A human reviewing the same data set may make inferences, observations, an intuition-based reflections that are beyond the capabilities of machine learning. This is especially true in situations that might be thought of as outliers.

Even with the advantage of improved efficiency, care needs to be taken to avoid overfitting.

Continuing Development

“. . . while we don’t see AI eliminating lawyer jobs, AI will fundamentally change what lawyers do and how they do it, in much the same way Westlaw and Lexis changed the way (and speed at which) traditional legal research was conducted.” – Above the Law

That’s an apt conclusion. Machine learning represents a powerful mode of AI within the legal industry. It will almost certain change how law is practiced. But it’s no replacement for human interaction, experience or intuition.

Part of our commitment to our clients includes following new developments, like the emerging field of machine learning and how best to apply it. That’s what sets Process Fusion apart from our competitors. Organizations that partner with us get top-tier support and the benefit of the latest available tools.

If you have process automation needs, we can help. Get in touch with us today to explore what we can do for you.