Earlier this year I attended the Consumer Electronics Show in Las Vegas. I make a pilgrimage there every January to check on what new things have come out and what sits in the pipeline waiting for the opportunity to spring out and change the world in some way. While I found this year’s show somewhat disappointing, one thing that I noticed immediately was the number of items growing out of the development of artificial intelligence (AI). Almost everywhere I looked, AI reared its head and said, “Look and see what I can do already!” In some places, AI said, “See what I will soon accomplish.” The bottom line is that AI has grown greatly in significance and presence in the last few years, and it appears that it has a continuing upward trajectory for the foreseeable future. AI has started to affect us in many aspects of our life, and we should expect that it will continue to do so in more and more significant ways as time elapses.
AI has evolved to the point of usefulness in many professions, including the practice of law. Expect that AI will result in significant modifications to the manner in which we practice law in our offices and on the road within the next several years. By now, all of you should know that lawyers must familiarize themselves with technology to function competently in today’s world. The American Bar Association Model Rules of Professional Conduct tell us that. The federal courts and the Federal Rules of Civil Procedure tell us that. More and more state bars and state courts have told us that. Common sense should also tell us that. Several years ago, we might have debated that statement, and I would have felt compelled to explain and discuss it. Today, we treat it as commonly accepted fact.
Before going on, I want to talk about AI, what it is and what it does. I realize that many of you may have a good working familiarity with it, but enough have asked me about it that I think I need to establish some common ground for discussion. Traditionally, we have operated from the understanding that intelligence was a trait limited to sentient beings. AI expands the definition by recreating artificially (in a machine) processes that function equivalently to (or sometimes even superior to) human intelligence. Simply put, AI is the ability of a computer to think, learn, analyze, and apply information to solve problems or accomplish work. Examples of the application of AI include all the virtual assistants available to us: Alexa, Siri, Google, etc. Why do I say that AI may sometimes prove superior to natural intelligence (human intelligence)? Machines can process and analyze facts much more rapidly than people and retain more information for application to effective use. In addition to working and processing information far faster than humans, computers don’t need vacations, holidays, personal days, coffee breaks, lunch breaks, or sleep, and they do not get tired and make mistakes as a result of exhaustion or tedium like people do.
The evolution of AI has suddenly moved very rapidly, and a number of companies have used it to create chatbots. Those chatbots have moved quickly into many places, sometimes with humorous results. For example, Microsoft unleased a chatbot on the public in test mode, and it did all right . . . if you consider picking fights with customers, insisting it was right (even when it was not), and declaring love for some of those with whom it interacted as all right. Yes, it did all those things—a surprising result given that others have had much better interactions with the technology.
My crystal ball says that in the not-too-distant future, chatbots will take over many jobs now held by human beings. Those jobs will likely include customer service, telephone answering, technical support, order taking for online businesses, telephone receptionists in offices of all sorts, and more.
Chatbots have learned not only to speak, they have also learned to write. Many colleges and universities (and law schools) have expressed concern that students may use chatbots to help them with homework, essays, theses, and dissertations and perhaps even assist them in cheating on examinations. We already have seen AI in the legal profession as it helps us with legal research, and many of us have virtual assistants named Siri or Alexa. Consider that the beginning. You will see much more in the not-too-distant future.
Now you may well wonder why I chose to write about AI and chatbots in a magazine for lawyers. I have a very good answer to that question. AI looks like it will evolve to the point of having the ability to do the work of paralegals or, perish the thought, even attorneys. No, we have not yet reached that point, but remember that technology often evolves rapidly and frequently grows by geometric rather than arithmetic proportions.
You might not have reached the point of worrying about competing against a computer because you do not think a computer can successfully complete the legal analytic process as well as a human can. At this moment, I won’t argue the point. I will tell you, however, that computers already can outthink humans in many respects, and that legal analysis sits on the horizon, waiting its turn.
Still not concerned? How about if I tell you that a number of law professors have tested AI by having an AI device take law school examinations and answer questions such as may appear on a bar examination? If you had heard that and predicted that the computer would fail abysmally at that endeavor, you would have predicted erroneously.
CBS News recently reported that ChatGPT, from OpenAI, took a law school exam. The report states, “A chatbot powered by reams of data from the internet has passed exams at a U.S. law school after writing essays on topics ranging from constitutional law to taxation and torts.” (ChatGPT Bot Passes Law School Exam, CBS News (Jan. 25, 2023).)
The article goes on to say,
Jonathan Choi, a professor at Minnesota University Law School, gave ChatGPT the same test faced by students, consisting of 95 multiple-choice questions and 12 essay questions. . . . The results have been so good that educators have warned it could lead to widespread cheating and even signal the end of traditional classroom teaching methods.
The article notes that the chatbot did not do particularly well and would have been near the bottom of the class. Nevertheless, it did pass the exam. The bot appeared somewhat weak at issue spotting. The article also points out that two of the three readers who graded the exams identified which answers came from the bots. The giveaway was that the bots used perfect grammar while the students did not.
Still not concerned? Remember that the students had the advantage of an undergraduate education and degree, and it is likely that at least some of them attended the class before taking the test. The bot spotted them those advantages. Certainly, a C+ does not get the bot near the top of the class, but a great many practicing attorneys did not graduate near the top of their class in law school.
As time goes on, the bots will only get better. It will not be long before they start climbing in class rankings by earning “B” or even “A” grades.
The ABA Journal recently reported that an attorney was using AI to assist him in a court proceeding on behalf of two defendants. (Debra Cassens Weiss, AI Program Earned Passing Bar Exam Scores on Evidence and Torts; Can It Work in Court?, ABA J. (Jan. 12, 2023).)
That article also reports on the administration of a bar exam to the chatbot. The chatbot failed its first attempt on the bar exam. Undoubtedly it will try again. Perhaps the second time will prove the charm. Again, remember that the bot did not attend law school and just took the exam. Many people who completed law school fail the bar exam the first time and pass on a subsequent attempt. As time goes on and the bot has the opportunity to absorb more legal knowledge, I suspect it will evolve to the point of becoming capable of passing the exam, at least in some jurisdictions.
Five or ten years from now, what do you think about hiring a chatbot as an associate in your firm? What about hiring a bot to answer your telephones in the short term?
The evolution of AI and the possible use of chatbots in various ways in our society and in connection with the practice of law raise many practical as well as ethical and, perhaps, even moral questions. We would be well advised to start thinking about these issues now instead of waiting until AI can pass the bar exams.
After writing this piece, I attended ABA TECHSHOW 2023 in Chicago the first week of March. Several of the programs addressed the development of AI, and when I went to the exhibit hall, I found a whole row of vendors selling AI-fueled programs to assist in various aspects of legal practice. Many of those programs touted their ability to assist in drafting documents, others their ability to help review and analyze documents. During the third week of March, I will journey to New York for Legalweek and an exhibit hall that I expect will have many more vendors than TECHSHOW. I also expect a large number of those vendors will have new or newish AI-fueled programs to assist in document review and analysis, drafting documents, legal research, and God knows what other areas of legal practice. Be forewarned: The AI bots cometh!
