In the end, after you've stripped away their six-figure degrees, their state bar memberships, and their proclivity for capitalizing Odd Words, lawyers are just another breed of knowledge worker. They're paid to research, analyze, write, and argue -- not unlike an academic, a journalist, or an accountant. So when software comes along that's smarter or more efficient at those tasks than a human with a JD, it spells trouble.
That's one of the issues the Wall Street Journal raised yesterday in an article on the ways computer algorithms are slowly replacing human eyes when it comes to handling certain pieces of large, high-stakes litigation. It focuses on a topic that is near and dear to the legal industry (and pretty much nobody else) known as discovery, which is the process where attorneys sort through troves of documents to find pieces of evidence that might be related to a lawsuit. While it might seem like a niche topic, what's going on in the field has big implications for anybody who earns their living dealing with information.
The discovery process is all about cognition, the ability of people to look at endless bails of info and separate the wheat from the chaff. For many years, it was also extremely profitable for law firms, which billed hundreds of dollars an hour for associates to glance at thousands upon thousands (if not millions) of documents, and note whether they might have some passing relevance to the case at hand. Those days are pretty much dead, gone thanks to cost-conscious clients and legal temp agencies which rent out attorneys for as little as $25-an-hour to do the grunt work. Some firms are still struggling to replace the profits they've lost as a result.
And now comes the rise of the machines -- or, more precisely, the search engines. For a while now, attorneys have employed manual keyword searches to sort through the gigabytes of information involved in these cases. But as the WSJ reports, more firms are beginning to use a technology known as "predictive coding," which essentially automates the process at one-tenth the cost. Recently, a magistrate judge in a major Virginia employment discrimination suit ruled that the defense could use predictive coding to sort through their own data, despite objections by the plaintiffs who worried it might not pick up all the relevant documents (Probably left unspoken here: plaintiffs in lawsuits also like to drive up the costs for defendants, in the hopes that it will encourage them to settle).
In truth, researchers have found predictive coding to be as accurate, if not more so, than the attorneys its replacing. As the WSJ noted:
Several studies have shown that predictive coding outperforms human reviewers, though by how much is unclear. A widely cited 2011 article in the Richmond Journal of Law and Technology analyzed research on document review and found that humans unearthed an average of about 60% of relevant documents, while predictive coding identified an average of 77%.
The research also showed that predictive coding was more precise, flagging fewer irrelevant documents than humans did.
"Human readers get tired and make mistakes. They get fatigued," says David Breau, an associate at law firm Sidley Austin LLP who has written about predictive coding.
Shorter version: There is now software that's smarter and more efficient at these tasks than a human with a JD. Not only that, but it's finally being given sanction by the courts, which would have the power to stop such a new technology in its tracks if they chose.
For the legal industry, this is a mixed blessing. The same way that robotics have created factories that need fewer workers, these programs will create firms that need fewer lawyers (even if it just means they're renting fewer temps). Firms that have already figured out how to prosper without the enormous margins they achieved by charging egregious fees for associates to perform menial labor will benefit. The partners will keep on doing the most valuable work, even as the bottom rungs of their firms shrink.
But what about the rest of us? We're now living in the age of Watson, everyone's favorite Jeopardy-playing computer, where intelligent software has become more skilled at recognizing and fetching relevant information than most people. It's precise enough to do our taxes. It's precise enough to satisfy a judge. And the people who benefit most from it are those at the top of an industry, the people like law firm partners who are paid to take information and shape it into a narrative.
This is how countries get more productive and more prosperous. As we learn to make goods and perform services more efficiently, their prices come down. The question is what new jobs replace the old ones that are lost in the churn. As of now, who knows what they'll be? Maybe in a few years there will be program I can ask.