The new robot lawyer will be better than the human one with the help of artificial intelligence (AI). The use of AI, although still in its infancy, is changing the legal profession profoundly.
What is AI?
Artificial intelligence is the term used when computers and machines are able to perform tasks that are characteristic of human intelligence. This includes understanding language, solving problems and making decisions based on patterns and data.
Computers can perform these tasks multiple times faster than any human. Because AI can scan and analyze large sets of data and learn from the data, it can see patterns that humans cannot.
Being able to learn tasks on their own, without human intervention, is known as machine learning. Machine learning is accomplished by giving data to the algorithms so they can learn and improve.
This is why the use of AI in the legal profession is a game-changer.
Legal Research and E-Discovery
The first areas of the legal profession that embraced the use of AI are legal research and document review, or e-discovery. Legal search programs like Westlaw and Lexus can compile hundreds of relevant cases in the blink of an eye, with great accuracy. Every decision handed down by judges in the country is scanned by these programs. Every word in each decision is a data point that is categorized and analyzed.
E-discovery can also benefit from AI. Large companies can create huge amounts of data in just one day. If a lawsuit requires examination of documents spanning 20 years, any law firm would need an army of document reviewers. With AI, however, this task can be accomplished in a fraction of the time, at a fraction of the cost and with greater accuracy.
To accomplish a high level of accuracy in e-discovery, predictive coding, a form of machine learning, is implemented. Predictive coding is very efficient at accurately locating and retaining pertinent documents in a data set. Predictive coding pulls out sample sets from the larger data set and codes them either as “responsive," a document is pertinent to the lawsuit, or “unresponsive.” With several sample sets coded in this fashion, the AI software can refine itself for better accuracy of predicting the responsiveness of future documents.
Predicting Legal Outcomes
One of the interesting but eerie uses of AI in the legal world is predicting the outcome of case or disputes. AI is becoming increasingly accurate in predicting how lawsuits will end. Because an algorithm can consider many more data points about a particular lawsuit or dispute, it can better predict the outcome of the case than humans. Further, by taking the human element out of the equation, the AI program is not biased by feelings or emotions.
For example, a London-based personal injury firm has created an algorithm, based upon 600 cases over a 12-month period, to determine the probable outcome of injury cases. With this type of tool available to firms, lawyers can easily and quickly assess the outcomes of new cases, and they can predict the likely settlement amount, chance of winning and the amount of litigation expenses that will be needed.
The fear in the legal community is that within the next 10 to 15 years, actual human lawyers will be replaced in certain areas of the law. Many companies have already developed AI programs that sort and review contracts, put them in different categories, and spot errors. These algorithms can perform these tasks quickly and error free, unlike human lawyers.
The Deloitte Insights, a group that tracks changes in the legal industry and tries to predict the future of the industry, indicated in its 2016 report that within the next 10 years, 39 percent of the jobs in the legal sector will be eliminated. That equates to well over 100,000 positions ended.
This process is already happening. Over the last 15 years, law firms in the divorce industry have replaced attorneys with AI programs.
Wevorce, a Boise, Idaho based company has designed a system that replaces divorce lawyers. The CEO and creator of Wevorce, Michelle Crosby, states that, “Divorce has really always been about lawyers doing things a certain way to get documents through. We’re starting to look at it differently.”
Wevorce charges $949 (excluding court filing fees) and takes an average of 30 days to compete a divorce. The process starts with both parties inputting each side’s best, most desired outcome, emphasizing their priorities and attempting to reach an amicable agreement on property distribution, child support, alimony and child custody.
Both parties answer guided questions. The AI algorithms guide the clients to different questions depending upon what answers were given. Then the program creates the necessary legal documents based on the information given. As Crosby explains, “I knew that so much of what lawyers do is really quite templated. I could take so many of those pieces and use software — not to remove human elements but to do the things that can be done a million times without people. You don’t need to pay someone $200 an hour to do these pieces.”
There are human experts and human customer support available, but much of the process is automated and the users can go at their own pace. As of January of 2016, the company has processed over $40 million in assets through its platform and is planning to expand its services through a series of partnerships, including one with the U.S. military.
Problems With Confidentiality
As with any new technology, law firms have to worry about confidentiality and data breaches. AI is no different. According to the American Bar Association, in 2017, 1 in 5 law firms either experienced a cyber attack or were hacked. This number was up 14 percent from the year prior.
However, forward thinkers in the AI legal world believe that the use of AI can be useful in preventing cyber attacks and data breaches. The use of self-learning algorithms can track attempts of security breaches and defeat these attacks before any data is stolen. Many think that law firms that rely upon outdated technology face a greater risk of breaching confidentiality by having their data stolen.
Good and Bad Data
Users of AI in the legal industry need to be aware of the quality of data they use. Since the effectiveness of AI depends on algorithms combing through huge data sets to find relevant patterns, it is very important to make sure the data is useful and relevant. Amassing quality data is the starting point in any use of an AI protocol.
As an increasing number of legal tasks are entrusted to machines, the advent of robot lawyers is becoming less of a fantasy. The integration between technology and the legal industry is not slowing down.