The grandfather of artificial document intelligence (ADI) is U.S. Magistrate Judge Andrew J. Peck, who authored the first-ever court decision approving the use of technology-assisted review in e-discovery (and was awarded the Champion of Technology award by Legaltech News). Before long, judges began mandating the use of e-discovery.
Technology-assisted review (TAR, aka CAR: computer-assisted review) represents the productive marriage of humans and technology. Essentially, the human reviewers train the technology, and then the technology uses what it has learned to go through potential electronically stored information (ESI): documents, emails, etc.
This is the same concept used by retail sites that collect your shopping behavior and provide recommendations of what else you might like to buy.
This technology has been applied to upload a collection of documents into a firm’s repository and allow users to train the systems about what information is relevant. This can be used for due diligence or when someone comes to the firm with new documents or a new client. The technology is able to review all of the documents and then identify relevant sections upon request, or proactively make recommendations to new documents as they are being drafted, based on recurring sections in other similar documents.
Millennials now make up more than a third of the U.S. workforce and as much as 21.5 percent of lawyers in the United States, according to the U.S. Bureau of Labor Statistics. Even if a law firm’s luddites wanted to ignore the incredible efficiency delivered by this technology, their younger attorneys wouldn’t let them.
Leveraging technology-driven efficiency is at the core of how Millennials work. If you don’t support them, you will handicap their productivity, and they will eventually figure out a way to do it on their own. It is critical that every firm provide the capabilities they want and the built-in protections that the firm requires.
Inevitably there will be inconsistencies generated by TAR, and also when a firm’s Millennial lawyers pull content from outside sources like Google. Anticipating and protecting against this is good for the security of the firm and provides a comfort level to the more senior lawyers who are cautious about new technology.
It also elevates your firm’s recruiting appeal to Millennials.
As predictive technology evolves, it will get better and better, but like all nonhuman intelligence, it will never perfectly replicate the human thinking process. Who among us has not received a recommendation on Amazon for something we have no interest in, or a song suggestion on Pandora we don’t like?
Unlike e-commerce sites that experience very little downside in suggesting another purchase, law firms have everything to lose if they fail to catch and correct an inaccurate prediction based on technology.
The idea of an engine that can offer a clause from your firm’s repository for use in a document you are actively working on is seductively convenient. However, the current challenge is that it could be adding ambiguity due to subtle differences in documents, or introducing a term that you haven’t defined.
This is where proofreading solutions help to maximize the efficiency benefits that TAR delivers, while minimizing the potential risk. While lawyers, chief information officers (CIO) and chief technology officers (CTO) admire the efficiency of proofreading software, the greatest service it provides is protecting the user from the risk of errors, inaccuracies and inconsistencies, including those generated by TAR. This is what partners worry about and what their clients care about.
If your reputation is damaged or you become entangled in costly litigation because of a mistake in a document, efficiency becomes irrelevant.
An internal repository goes a long way toward minimizing exposure. Ideally you want to use only content that you trust that is consistent with the way the rest of your documents are written.
We are seeing progressive firms push for more sophisticated comparisons and information. Lawyers want the assurance of knowing that nothing is missing from their documents. They also want confirmation that the language is consistent: 1) throughout a given document, 2) with similar documents created for the same client, and 3) with all documents created by their firm.
A likely evolution of this search functionality is the ability to create a dynamic template that pulls content from all the documents a lawyer has “liked.” Imagine the time savings delivered by software that automatically generates a new contract based on a lawyer’s favorite language from every similar document they have ever created!
What we are seeing now is the ability to set up stipulations within internal repositories regarding what should be made available for use and what shouldn’t, e.g. documents that haven’t been accessed or modified within five years, or were created by someone no longer at the firm. Did a partner work on it? Who drafted it? When was it last proofread? This information is all in the metadata.
Applied in the right way, metadata can be leveraged internally for analytics and AI, and it can make a firm’s repository easier to navigate and more reliable. Just don’t forget to scrub any documents that are going to leave the firm!
One of the more creative applications on the horizon provides insights into a judge’s ruling by sharing document language from winning cases they’ve ruled on in the past. These smart edits or corrections could pop up on a lawyer’s screen during the drafting process to offer favorable language for a certain judge or red-flag words that might be unpopular based on their previous rulings.
Now that the gates have been opened, there is no turning back. The mindset of lawyers has shifted, and they are now asking: What else can technology do to save time?