Across all industries, more than a third of businesses have adopted artificial intelligence (AI) into their operations. In finance departments, marketing and sales, research and development — a wide variety of business functions are making use of legal AI to streamline operations and do more with less.
The exception is the legal department. According to Thomson Reuters, only 21 percent of legal departments believed mainstream adoption of AI will occur within five years, while 39 percent predicted it will be within 10 years, and 37 percent believed it will take more than 10 years. Meanwhile, a Deloitte report argued that up to 39 percent of legal jobs could wind up being automated away within 10 years.
So why the disconnect?
Until recently, the legal industry has been fairly resilient against the forces of technological change, but the advances in modern AI technology is quickly changing that. As other business functions quickly embrace AI, many legal departments may fall out of step with their larger organizations, struggling to keep pace with a business powered by real-time insights and predictive analytics. That’s why learning how AI works and how it will impact the legal profession is key.
How AI Works
Once we dive into the fundamentals, it’s clear that AI and the legal profession have a significant amount of overlap. Both fields rely on historical evidence to extrapolate rules to apply to novel scenarios. While there are many techniques and fields of study within AI, legal professionals will most likely work with just three.
1. Machine Learning
For many, machine learning and AI are synonymous with one another. Where AI simply refers to making machines behave and perform tasks that, until recently, were believed to require human intelligence, machine learning refers specifically to the development of algorithms that improve with experience.
Frequently, machine learning relies on data sets where the internal relationships and dependencies of the data have been pre-identified. For example, a legal professional might provide a machine learning solution a data set of contracts that have had non-standard clauses and language pre-identified. After ingesting that data set, the machine learning solution would be able to identify such clauses in new contracts.
Obviously, this is a major simplification, but the important thing to understand is that any machine learning solution is going to require training by a legal professional. Thus, you don’t have to worry about a robot taking your job — at least not for a long time.
2. Optical Character Recognition (OCR)
OCR is a technology based on machine learning, but it doesn’t require any input from a legal professional. The entire purpose of OCR is to recognize letters and words from images so that they can be digitized.
For legal professionals, OCR is both a significant time-saver and a foundational solution. Handwritten notes and scanned court documents can be translated into a machine-readable format, ensuring that they can be easily shared with colleagues, stored and — most importantly — ingested by other AI solutions.
3. Natural Language Processing (NLP)
NLP enables machines to understand human language. When searching for an answer to a query, for instance, NLP would be able to identify more than just a given keyword, but rather understand the meaning of your question and the different ways in which it could be answered.
To a limited extent, we’re all familiar with NLP’s capabilities through Google. But in the legal profession, with its hyperfocus on the perfectly unambiguous use of language, advanced purpose-built solutions are needed.
How You’ll Use AI
We’re only scratching the surface. AI is — unsurprisingly — a deep topic that you could happily dedicate a lifetime to studying and learning. But as a legal professional, there’s already enough on your plate without adding the burden of learning advanced computer science. The point of AI is to make legal professionals’ work easier; let’s dive into how you’ll be using AI to do just that.
1. Pre-signature contract review
It’s a simple fact that the more people are involved in a project, the longer it will take; reviewing contracts between your and another organization can involve a dizzying array of stakeholders, each with different requirements and competing interests. The process of passing contracts back and forth, redlining contentious language, revising and re-reviewing isn’t quick. Not to mention that handling different versions of a document increases the odds of introducing error.
AI solutions that make use of NLP can ingest draft contracts and quickly identify problematic, high-risk language. For a single, straightforward contract, such solutions aren’t likely to result in significant time savings. But for organizations that have high volumes of contracts to review or highly complex contracts to review, they can give a business the edge that comes with moving fast.
2. Post-signature contract analytics
Any single contract has a few straightforward characteristics to track: renewal dates, obligations, individual terms, relevant stakeholders, etc. But multiply these requirements over an entire contract portfolio that might number in the thousands or tens of thousands, and it’s clear that a better method of tracking contractual obligations and identifying contracts of interest is needed.
This requires first digitizing any paper contracts or poorly scanned contracts using OCR. Afterwards, NLP-based solutions can identify key data points throughout your portfolio. With this kind of AI technology, you can quickly determine which contracts have, for example, a force majeure clause, which contracts are up for renewal this quarter or which contracts offer attractive options that your organization isn’t currently exercising.
Document sets have grown too large for even a team of highly trained attorneys to review in an expedient manner. Using a machine-learning technique called predictive coding, legal professionals can pick a subsection of the total document set and identify the aspects of individual documents that make them relevant or irrelevant. Using this coded subset, a machine learning algorithm can “learn” what characteristics make a document relevant to your matter. With this technique, once impossibly large document sets can be culled down to the most likely candidates.
4. Litigation prediction
While there will always be an element of chance, many of the factors that go into determining whether a given case will have a favorable outcome are known and quantifiable. By ingesting data on the facts of the matter, previous case law, the judge’s previous rulings and jurisdiction, machine learning algorithms can estimate the likelihood of a successful outcome. In fact, several studies have already proven the concept by correctly predicting the outcome of several Supreme Court cases, and several start-ups have appeared that offer litigation prediction platforms.
5. Legal research
Where the bulk of legal research was once an extremely manual process handled by junior lawyers, today there are better, more efficient options. Modern legal research platforms leverage NLP to go beyond merely matching keywords. Platforms like ROSS Intelligence, Westlaw Edge and others are capable of identifying the semantic relationships between cases, surfacing highly relevant precedents from question-based search.
How to gain an edge with AI
The legal profession isn’t known for its agility, which, when it comes to adopting AI, is a double-edged sword. On the one hand, it will be difficult to get buy-in at your organization to invest in potentially transformative but relatively new technologies. On the other, those organizations that move fast stand to gain a serious advantage.
The key to falling under the latter camp rather than the former is to build an understanding of how these technologies work and how they can be applied in your organization. Read our article, How Data Is Shaping Corporate Legal Team Decisions, to fully flesh out your understanding of AI in the legal department.