[Note: this is a summary version of a four-part series I wrote for Thomson Reuters earlier this year with some key updates and new sections. You can find the full series by clicking here]
Some of the best scenes in the television show Star Trek (the original version) involve the crew members asking the computer a question and the computer spitting out the answer in the form of a conversation. I was utterly amazed by this and, of course, wanted my own computer that would “answer” my questions. Alas, I was ahead of my time. But, I was recently reading an article comparing the Google Home with the Amazon Echo, two devices that allow you to ask questions and get an answer. From a computer. Just like Star Trek! This brought back two important memories from the TV show: 1) intelligent computers that can talk and answer questions; and 2) never be the crewman in the red shirt. I always lived by the latter and now, it dawned on me, I could start to live by the former too.
1. What is AI? The term “Artificial Intelligence” can be a bit misleading, at least when it comes to application in the legal field. Perhaps a better description, and one that is catching on, is “cognitive computing.” This means computers learning how to complete tasks traditionally done by people, where the focus is on looking for patterns in data, testing the data, and finding/providing results. Put another way, “a research assistant” who can sift through the dreck and tell you what it found. Why is this important? Because, according to IBM, 2.5 quintillion bytes of data are being generated every day. In case you’re not up to date on “quint,” that’s 2,500,000,000,000,000,000 bytes. Every day. The ability of any human to review and comprehend that level of data without help is impossible.
2. How does it work? AI is the science of teaching computers how to learn, reason, perceive, infer, communicate and make decisions like human’s do. The initial goal is called machine learning, where a computer begins to make decisions with minimal programming. Instead of manually writing rules for how the computer should interpret a set of data, machine learning algorithms (i.e., instructions for solving particular problems) allow the computer to determine the rules itself. Ultimately, computers become better at their tasks with experience. Cognitive computing requires three core processes: 1) gather information, 2) analyze and try to understand the information, and 3) make decisions based on this understanding. As all lawyers know from experience, this process is iterative and we become better the more times we undertake the task – especially if we are corrected and guided in our work by someone more experienced (just like being a young associate at a law firm). It works exactly the same way with AI.
3. Why does it matter? AI has the potential to solve two huge problems all legal departments face: 1) lack of budget and 2) lack of manpower. If the use of AI can provide less expensive legal services (either used internally or purchased through a law firm), the value is immediately apparent. If the use of AI can free up current staff from spending time on transactional-tasks, the value (if not immediately apparent) is exponentially greater than merely paying less for legal services. Freeing up attorney time gives you and your business clients more access to better and more involved legal services (i.e., attorneys who can dedicate more time to thinking through problems and advising clients). The combination of these benefits will allow in-house legal departments to actually deliver on the old CEO/CFO demand of “doing more with less.”
4. How can I use AI in my Legal Department? The AI use most in-house lawyers are familiar with involves e-discovery. Originally, the AI use here was pretty simple – looking for key words in megabytes of data. Doing this saved lawyers an amazing amount of time (and money) that would otherwise be spent paying lawyers to find relevant documents. Later, AI was used to eliminate duplicate documents and connect strings of emails, again doing in minutes that would take days if done by people. Finally, AI is now capable of searching discovery documents for context, concepts, and tone with what is known as “predictive coding,” going far beyond simple “keyword searches.” You still need lawyers, but you no longer need an army of them.
5. What else can AI do? I think the biggest impact comes from how AI is (or will be) used by in-house legal departments. This is true because the incentive for in-house legal departments to find the lowest cost way to do things is greater than the incentive for law firms to find ways to offer lower priced services (i.e., quantity of billable hours vs. quality of a billable hour). Here is a list of some of the things AI can do (or is coming) for in-house legal departments that will disrupt the legal market for years to come:
• Due diligence reviews. Due diligence reviews for a corporate transaction typically involve a bunch of lawyers going through documents (hard copy or in an e-room) looking for litigation issues, key contract clauses (e.g., change of control, assignment, etc.), corporate governance, intellectual property, etc. Generally it takes many hands (usually outside counsel) and many hours/days to complete. AI can do this in a fraction of the time.
• Prepare contracts. The “Holy Grail” for in-house lawyers who draft contracts is the ability to create and use a form agreement, i.e., one that has standard terms and conditions and allows limited changes/customization. Form contracts are huge time savers and allow the company to have a consistent set of agreements. There are AI tools that can create contracts, using whatever set of parameters the legal department feels important. Moreover, the tool can be set up as “self-service” for clients, i.e., the client can log onto the system, select the type of contract they need, enter in a few variables, and the system will produce a standard form agreement ready to go.
• Contract management. Historically, contract management is done manually. Someone either creates a “spreadsheet” and tracks everything by hand, or enters the data manually into a system that manages the key terms and dates automatically. AI has progressed to the point where the entry of the key information (parties), terms, dates and other information can be done by the computer – without the need for human intervention (other than initial set up and fine tuning). In fact, contract management — including signature process — can be managed as part of the same tools that create the contracts, meaning an automated process from start to signature to storage.
• Legal spend/Legal operations analysis. Most Legal Departments are moving to e-billing systems. There is a wealth of information in the e-billing system but many in-house lawyers are not good at extracting the information in a useful manner. AI provides the capability to analyze what work was done, how it aligns with other work done by that firm, how the work/efficiency compares with work provided by other firms, and how the work/efficiency compares to the market generally. Imagine being armed with this information the next time you want to discuss billing rates with your law firms.
• Litigation analysis. There is an amazing amount of data in the US court system’s public records. Opinions and orders of courts, jury verdicts, and other valuable information is generally fully available. Wouldn’t it be great to search all of that data and use it to predict the outcome of litigation? AI is already providing a solution here as well. It will soon be possible (or perhaps it’s already possible) to compare the facts of your case to other cases already decided by a court (or courts) and get a predictor of how your case will fare.
• “Wrong Doing” detection. You may remember the 2002 Tom Cruise movie called Minority Report. Mr. Cruise played a member of the Precrime Police, a group of police men and women who stop crimes before they occurred because they had access to information that told them the crimes were about to happen. Granted, the information was provided by three clairvoyant mutant humans floating in a milk bath but let’s ignore that for now. Taking the concept of predictive coding further, it is now possible to utilize AI to search company records (documents, emails, and even unstructured data) to detect “bad behavior” before it can bubble to the surface. AI is being used to sniff out bribery, fraud, compliance issues, even “potential” litigation – all based on the content of the company’s own documents and data. AI can summarize conversations, the ideas discussed, note the frequency of the communications, and even the “mood” of the speakers.
• Legal Research. In-house lawyers generally short-change the research process because they don’t want to spend the time or money to do a complete job, or they overpay either in terms of time they spend on a relatively mundane task or by paying a law firm to flail away on the question. AI tools will soon allow you ask legal questions in “plain English” and get an answer back. An answer that includes research of regulations, case law, secondary sources, etc. Moreover, it may be possible to use AI as a “talking” FAQ service (chatbot) that can answer basic legal, HR, and compliance questions from your in-house clients, yet be “smart” enough to know when to defer the answer to a live attorney.
6. What’s the cost? As with many new and revolutionary technologies, a significant issue revolves around the cost of getting the AI resources up and running. At the 2017 Legalweek Conference, the chairman of the board of the Corporate Legal Operations Consortium, Connie Brenton, cautioned attendees about the reality of cost vs. the novelty of AI. In an article written about the 2017 Legalweek Conference, Rhys Dipshan used Cisco as an example of the current high cost of implementing AI on a large, enterprise level. In that story, he reported on Cisco’s implementation of AI applications in the virtual assistant and contracting space. Cisco estimated the cost for 500 users to be around $250,000 plus the need to hire two-and-a-half full-time employees internally to implement the tools for a basic system.
The ROI of AI will be heavily fact-dependent for each organization based on what AI tools the department wants to utilize, the amount of customization needed, the number of users, and so forth. But of course, over time, as adoption increases and Moore’s Law takes over, the power of AI will increase and the costs will come down, probably dramatically within the near future
7. Don’t fear the robots. AI will impact jobs in the legal industry and I see three things happening:
• Some legal jobs will be eliminated, e.g., those which involve the sole task of searching documents or other databases for information and coding that information are most at risk.
• Jobs will be created, including managing and developing AI (legal engineers), writing algorithms for AI, and reviewing AI-assisted work product (because lawyers can never concede the final say or the provision of legal advice to AI).
• Most lawyers will be freed from the mundane task of data gathering for the value-added task of analyzing results, thinking, and advising their clients. These are roles that will always require the human touch. AI will just help lawyers do this better, faster and more cost effectively.
Matthew Francke wrote about several reasons why people should/will never entrust a computer to provide legal advice:
• Judgment is a skill. Good lawyers (like good doctors) will – and need to – take the time to understand their clients, what they are seeking, and why. Judgment is needed as part of providing any legal advice. Computers cannot do this in the same way humans can.
• The answers to legal questions are not always black and white. If you’ve practiced for any length of time, you already know that there is no Big Book of Law that contains all the answers to legal questions. It is rare that many legal issues can be solved by simply looking up a provision in the law. Instead, most of the law is built on precedent where every case had someone arguing the exact opposite of how the case was decided (and it may have been a close call, hinging on one key or obscure fact or other twist). Meaning, everything needs to be put into context, something computers are unlikely to handle.
• Successful people don’t use lawyers less; they use them more. Successful business people want to talk to a person, a counselor. It’s the personal relationship that cannot be replaced by a computer.
• The desire for engineered outcomes. It’s human nature to try to engineer the answer you want to get from a question. Asking “Can I do [x]?” vs. “Should I do [x]?” How a question is framed usually puts things in the best possible light to get the desired answer. A human can look at the circumstances, the body language and other intangibles, along with teasing out the relevant details so that you get not only the answer but help with understanding it, so that you can accept it and implement it.
8. This is all ethical, right? Lawyers must still stay involved and, ultimately, take whatever output from AI resources and apply good old-fashioned human legal analysis to come up with the actual legal advice. Not only is this just a good idea, it’s also what is required under the professional ethical canons – certainly here in the U.S. According to Wendy Chang, a member of the ABA’s Standing Committee on Ethics and Professional Responsibility, “In using technology, lawyers must understand the technology that they are using to assure themselves they are doing so in a way that complies with their ethical obligations – and that the advice the client receives is the result of the lawyer’s independent judgment.” In short, lawyers cannot ignore their ethical obligations and abdicate to technology.
9. My Vision. This is my idea for using AI to build an awesome tool for in-house lawyers. I call my computer “Holmes” because a) it’s a cross between Oliver Wendell Holmes and Sherlock Holmes, and b) it sounds cool – way better than calling it “Legal Computer.” My idea (perhaps not that all original) is for a dedicated desktop box and you can ask it complicated legal research and related questions and get useable answers. It will give you the answer in either spoken form or in an email/Word document. I see “Holmes” being able to handle two types of inquiries. The first is “basic legal research.” The second I call “deep research.” Here is a sample conversation with Holmes regarding some simple legal research:
Me: “Good morning Holmes. I need you to help me with some legal research.”
Holmes: “Good morning, Sterling. How may I help you?”
Me: “I need to know when our answer to the complaint in the Anderson matter is due.”
Holmes: “One moment please. Are you asking about the Anderson state court case or the Anderson federal court case?”
Me: “The federal case.”
Holmes: “We served with process on October 1, 2016. Our response is due 21 days later, October 22. However, since October 22 is a Saturday, our response will be due on Monday, October 24.”
Me: “Thank you Holmes. Please send me your answer as an email and prepare a short letter to opposing counsel requesting an additional 10 days to respond.”
Holmes: “I will send it to you shortly.”
Me: “Oh, and one other thing. Please send me the controlling case setting out the standard for granting summary judgment in the 5th Circuit.”
Holmes: “It will be my pleasure.”
You can instantly see what a time saver an AI product like “Holmes” would be for in-house counsel. And if you can tie it to a service like Lexis/Nexis or Practical Law you can even ask it to send you draft contracts, checklists, sample memos, research articles, etc.
My other idea for “Holmes” involves “deep research.” By this I mean to use the capability of AI to a) search thousands and thousands of documents for “wrong doing,” b) look through gigabytes of unstructured records and data for patterns, tone, and context, and c) access the entire internet and other public data sources (e.g., court and police records) for information I might need to know about. My vision goes like this:
Me: “Good morning Holmes.”
Holmes: “Good morning, Sterling. How may I help you?”
Me: “Is there anything I need to be aware of today?”
Holmes: “One moment, sir.” [Pause of thirty seconds or less]
Holmes: “Yes, Sterling, several things you need to keep your eye on today.” First, it appears that there may be a problem in our Brazil office that might involve bribing a foreign official. I will prepare a report for you summarizing the e-mail traffic and other documents and provide the names of the employees involved. Second, there may be someone forwarding trade secret information to a personal email account with the intention of leaving the company and starting a competing business. I will send that report to you shortly. Third, I have found reports on the Internet that the Attorney General of Kentucky may be working on an investigation that could involve our company and several of our competitors. I will collect that information for you as well. Finally, it is your wife’s birthday tomorrow. Do not forget.
Me: “Thank you Holmes, especially for the reminder about my wife’s birthday. Can you make reservations for two at Bleu Maison for tomorrow night around 7:00 pm and when you send me the report about the trade secrets, please copy Julie Smith as I will want her to help me with looking into that. Thank you.”
Holmes: “Will do, sir. Anything else?”
Me: “Not right now.”
When the capability for something like this arrives – i.e., warn in-house counsel of trouble brewing in the bowels of the company – you have an extremely valuable tool, one that can save in-house lawyers incredible amounts of time and money, and potentially prevent big legal problems like bribery issues, sexual harassment, fraud, employee theft, etc. Further, a tool like “Holmes” can scan the Internet and public records and look for any potential issues that might affect your company — the perfect marriage of law and Star Trek.
Of course, there are some potential problems here. There are employee privacy issues to work through and at a minimum, you’ll need policies to inform employees that information on the company’s servers is subject to search by the company at any time (and even then something like this might not comport with the laws of every country). Second, you might get overwhelmed with information. Third, who’s guarding the guardians? If “Holmes” is searching every email and all of the company’s data, what happens if it reports something sensitive but not necessarily a company issue, e.g., the CEO is having an affair but not with a company employee? What do I do when that type of information comes across my desk and who is making sure I don’t use that information improperly or at all? Finally, how comfortable am I that that “Holmes” got the answer right? Is there the potential for “malpractice”? And in terms of drafting, is “good enough” good enough? These are all issues that will need to be worked out and there will likely need to be updates and changes to codes of professional responsibility that provide ethical guidance to lawyers (in-house and outside) who may now have access to seemingly infinite information and the real capacity to “review” all of that data in a useful timeframe.
10. What should I do next? The future use of AI in the legal realm is both exciting and overwhelming. Like with any new technology, there will be early adopters and the costs will be high at first and then drop drastically as the technology becomes widely used. Here are my recommendations about what to do now with respect to AI:
• Partner with innovative outside counsel. Find out which of my outside law firms are using AI or intend to use AI and volunteer to be part of any experimental use of the product – in exchange for a low cost point.
• Identify an internal AI champion. Put someone in charge of AI for the department, i.e., an AI “champion.” Their role would be to learn all there is to learn about AI and how it might best be used in the department.
• Keep learning and researching. Find a continuing legal education program on AI and/or attend a trade conference on legal technology and dive deeper into the opportunities around AI to see firsthand what innovations and developments are coming in the form of AI. Check out the International Association for Artificial Intelligence and Law.
• Take baby steps. Find one use for AI in your legal department and start there. Don’t feel you have to go all-in on AI. It’s perfectly fine to start small and take your time. Look at all of the processes and work you do today and is there something that is a good candidate for AI assistance?
• Create an AI village. Make a good friend in your company’s IT department, someone with an interest in AI who can help procure resources and be supportive of Legal’s use of AI. Make sure the colleague is part of your team dedicated to AI issues and that you bring them in on the key decisions. The more vested they are in the success of your use of AI, the better off you will be.
• Resources. Here are articles you can read to get up to speed on Artificial Intelligence in the legal world:
- "How Artificial Intelligence is Transforming the Legal Profession,” ABA Journal, posted April 1, 2016 (Julie Sobowale).
- “Cognitive Computing: Transforming Knowledge Work,” Thomson Reuters AnswersOn Blog, posted January 24, 2017 (Dr. Khalid Al-Kofahi).
- “Artificial Intelligence (AI) in Law Departments: Opportunities,” LinkedIn Pulse, posted October 5, 2016 (Peter Krakaur).
- “Rise of the Machine – Artificial Intelligence in the Practice of Law,” ABA Litigation News, Winter 2017, Vol 42 No. 2 (Daniel Wittenberg).
- “Will Computers Replace Lawyers?” (Andrew Arruda)
- “Artificial Intelligence Systems and the Law,” Peer to Peer, Summer 2016 (Andrew Arruda)
Sorry for the length of this post but there is just so much going on right now in the field of AI and law. Artificial Intelligence just may well be the final frontier in terms of how legal services are utilized and provided. As in-house counsel, don’t run away from it and don’t ignore it. Rather, embrace it as, ultimately, it will allow you to do thing things lawyers love to do: thinking, analyzing, and counseling, while leaving the “grunt” work to the computer. If you’re interested, I will be on a great panel speaking about AI at the ABA’s National Legal Malpractice Conference on September 14 in Colorado Springs. Come up and say “hi” if you’re there.
Ten Things You Need to Know as In-House Counsel: Practical Advice and Successful Strategies is now available for sale. Described by the American Bar Association as “The one book all in-house counsel need to own!” Click here for details on how to order. Perfect for your library, or as a gift to clients or members of the legal department (or your next legal offsite).
If you find this blog useful, please click “follow” in the top right and you will get all new editions emailed to you directly. “Ten Things” is not legal advice or legal opinion and represents my views only. It is intended to provide practical tips and references to the busy in-house practitioner and other readers. If you have questions or comments, please contact me at either firstname.lastname@example.org.
My first book, “The Evolution of Professional Football,” is available for sale on Amazon and at www.SterlingMillerBooks.com.
 “Why You Want a Lawyer (and Not a Robot),” Best Hooper website, visited February 19, 2017 (Matthew Francke).