What “Teaching Legal Tech” Could Mean

The balance between going too far and not far enough

Ken Grady
The Algorithmic Society
11 min readMay 1, 2018


I started working in the legal industry in 1979, what many today would consider the dark ages. I was a paralegal at a large law firm working on a large lawsuit. Because I had worked for a year before joining the firm, I had the responsibility of coordinating the many paralegals working on the case.

Those were the days of manual discovery. The case involved a nuclear power plant and its fuel. The paralegals went through our client’s files. That meant looking through engineers’ and support staff’s files. We looked for documents responsive to the voluminous discovery request, hand-stamped each page of a responsive document, and logged the documents’ particulars into a massive index. It was all manual labor. To create the index, we dictated the necessary information (to, from, date, number of pages, topic) and secretaries transcribed the tapes. Next, we photocopied the index pages, cut the copies into strips (one per entry), and rearranged the strips into date order. That way, we had an index by “Bates” number and an index by date.

Paralegals assisted with witness files. We would locate all the documents to, from, or mentioning “Mr. X”, and work with the lawyers to cull that set to the relevant documents. We would create topic files (the story of how Y was approved). We, of course, had the smoking gun files.

Given the size of the case, the discovery files were massive. Most of the firm’s paralegals worked in a building across the street from the law firm’s main office in Chicago (known as the TransUnion Building or the TUB). Lining our office walls and the hallways, four-drawer vertical file cabinets — hundreds of them — formed our own steel grand canyon. Keeping track of the paper on our lawsuit alone was a full-time job for a dozen or so paralegals. Our client was one of the largest corporations in the world and the risk from the lawsuit was massive, so it was a classic example of “leave no stone unturned” litigation in the days before computers.

Falling Into The Coding Trap

A few years later, I was in law school and the personal computer had arrived. Along with it came amazing software such as the first generation of the database program, dBase I. This software became the leading database program for years. I was working at a boutique law firm part-time and our litigation opponents were large law firms. That meant we were physically outnumbered on every case.

To help level the playing field, I started programming in dBase creating a database program for discovery documents. Instead of dictating information about a relevant document, we would enter it directly into the database. We could then search the database and organize the output. Within minutes, we could locate all the documents mentioning “Mr. X” between date 1 and date 2. We could create privileged document logs at the push of a button. What took days and even weeks in 1979, took minutes in 1983.

I had done computer programming starting in middle school, so playing with the new personal computers was fun for me. But, you could have challenged me on whether it was the best use of time for a junior lawyer. The firm might have been better off if I had stuck to law and it had hired a programmer to create the database. As lawyers, we could use the output but we didn’t need to know the intricacies of how the program worked.

The Modern Legal Tech Debate

The debate rages today over what level of tech training law schools should provide to law students. Others debate how tech knowledgeable practicing lawyers must become. At one end, some law schools are teaching how to code artificial intelligence and machine learning. At the other end, many law schools are ignoring tech. We also have schools teaching specific legal tech software and others teaching basic skills. Among those already practicing, we have a huge universe of lawyers who know nothing about modern tech.

One approach to this dilemma is to simply shrug our shoulders and say “let the market decide.” Students can opt to go to a school that gets into the nitty gritty or stick with a school that says tech is for your employer to teach you. Clients will opt for tech savvy or tech ignorant lawyers. Eventually, clients and disciplinary actions will weed out those who fail to know enough about tech.

Many schools face the challenge of not having qualified professors on the faculty. This has led to the use of adjunct professors to teach many tech and other “New Law” courses. Adjunct professors (and I am one) can do a great job and bring real world examples into the classroom. But, as Sarah Glassmeyer has argued, adjunct professors are not the equivalent of tenure-track or tenured professors. If legal tech is central to the future of the profession, shouldn’t law schools step up and hire tenured faculty to do teaching and research in the area? (Yes, and some are: see Michigan State University College of Law’s hire of Professor Carla Reyes, who will start teaching in the Fall 2018 semester. Vermont Law School is another example.)

We still face the fundamental question: how much tech is enough? For example, Gabriel Teninbaum and David Colarusso teach tech at Suffolk Law. Colarusso teaches machine learning to law students as part of Suffolk Law School’s Legal Technology and Innovation Lab. The students use ML on real life projects. But another approach would be to teach the theory of ML, risks associated with ML (e.g., bias), and run through examples of how ML is being used, but not get into coding.

I lean towards not getting into the coding for several reasons (though I greatly respect what Gabe, David, and other tech professors in law schools are doing). First, it has been my experience that state-of-the-art tech quickly outpaces all but the most dedicated users. As I noted at the beginning, I began programming in dBase I. I kept up with dBase until somewhere around dBase III-Plus/dBase IV. At that point, the time required to keep up exceeded the time I had available to spend on programming versus law. At the pace tech moves, you either find yourself dedicating too much time to tech and not enough to law, or the reverse. It is hard to stay up to speed in both areas.

I had the same experience with Microsoft Windows. Initially, I kept up with the software doing things like going into the software registry to make changes. But at some point (I don’t recall exactly when) the complexity of the software overtook the time I had available to keep up with it.

Today, AI and ML fit into that same mold. As few as five years ago, one could keep up with developments in the area. But today, ML has become an exercise in math. To stay current, you need to read the papers released daily and become familiar with a number of tools and techniques. All of that can be done, but it takes time and that time has to come from somewhere — for a lawyer, from practicing law. There will come a time when the software is so powerful you can simply point it to a database and the software will do the rest. We aren’t at that point today. So, if you don’t plan to be an ML programmer applying the current methodologies to law, should you spend time learning ML?

I think law schools would be better off teaching law students basic concepts and skills. We did this in my Artificial Intelligence and Law course. Students needed to know that AI was essentially pattern recognition software. We ran through examples and applications. We talked about bias in creating the software and how that bias could affect tools lawyers used as well as applications lawyers might encounter. Most of the time, we talked about how the law could or should adjust to handle AI and ML. The students needed to know about AI, its capabilities and risks. But, they did not have to know how to code.

Do You Need To Learn To Code To Be A Better Lawyer

There is another argument raised in favor of teaching lawyers, and pretty much anyone, to code. Coding, they say, is an exercise in applied logic. As you learn how to code you learn how to be precise and thorough in your thinking. The computer cannot guess what you want next nor is it good at taking an instruction and saying “what you really mean is this...” Computers are literal. So, your instructions must be precise and in the correct sequence. You also must think through the possible outcomes and plan for all of them (the computer-world equivalent and the fertile octogenarian in law).

I do not dispute this argument, but I also do not think coding is the only way to get there. Instead, I would argue that law students should be challenged to use precise thinking as they work through legal problems. If they are not, the answer isn’t to teach them programming; the answer is to do a better job teaching them how to think analytically. Law has plenty of opportunities for thorough critical analysis without dragging coding into the game.

The Time Challenge

In some ways, the debate comes down to a time challenge. We have three years to teach law students what they need to know to hit the streets “practice ready”. Some want to shrink that to two years, to save money and eliminate some law school boredom. Either way, we need to decide how to make the best use of the time a student spends in law school. Exposing them to tools like tech, project management, process improvement, and data analytics will help them as they learn how to manage matters.

I am a big advocate of learning how to use tools essential to our profession. Law students should know how to use word processors, spreadsheets, presentation software, research software, and other basic software tools commonly used by lawyers. This, to me, is the same as teaching medical students how to use tools essential to their profession. But, to go the next step and teach coding is overkill, given the time and resources available. There will be some law students who decide a career involving coding and building applications to legal services is preferable to practicing law. That is fine. But, law schools should not be the places teaching coding.

To extend this thought, I also do not think law schools should teach specific software outside of what is commonly used in the profession. First, to me, this is an elitist and time sucking use of resources. Software companies will make their products available to law students free of charge. But most law students (remember, ~40% go solo or join firms of 10 or fewer attorneys according to NALP statistics) will not be able to afford the software when they graduate. Learning how to use, for example, a project management package instead of learning project management basics is a double negative. Students lack project management skills and valuable law school time is spent working with software they will never use.

I understand that teaching trendy software is fun, for students and professors. But, giving dessert to children can be more fun than having them eat a well-balanced meal. The well-balanced meal, however, teaches children what they should do to build a lifetime of health. Teaching basic skills teaches law students the foundation for a lifetime of learning, that can be applied regardless of what software comes into vogue.

This is different than teaching practicing lawyers specific software. If a practicing lawyers has the wherewithal to license software for use in her practice, then by all means she should get trained in how to use the software. But then we are matching training with specific users.

It Isn’t Just Academic

Law schools face a difficult environment today. The traditional approach of teaching legal theory and letting practice skills come after law school has come under a lot of pressure. Large law firms, which at one time hired as many as 40% of all law school graduates in the U.S., today hire under 20%. At the same time, about 40% of all law school graduates in 2016 went to solo firms or firms with 10 or fewer lawyers. The pressure for “practice ready” lawyers is increasing. Beyond legal theory, it can be difficult to identify precisely what tech skills new lawyers will need. Difficult as it may be, law schools need to do more.

I tried an experiment in one of my classes last year. The idea was to use free software for all the computer tasks needed in the class. Students had to write two papers, so they used Google Docs. It did not go well.

Out of a class of about 40 students, the majority had difficulty writing their papers in Google Docs and submitting them to me. The students clearly lacked basic tech skills (and students had trouble with the second paper even though they had some training with the first paper). The students also had trouble with other basic (free) software that would help them in their practices. In sum, the students were not “practice ready” when it came to basic software, even though most were from the “tech savvy” generation.

Integration Instead Of Add-On

I think there is another solution to this challenge that would work better than the current approaches, but it would take cooperative faculties. By incorporating tech (and project management, and process improvement, etc.) into law school courses, we could avoid the question of whether to add separate classes. If these methodologies become part of the learning experience, students will see them as part of how to practice law. They should be integrated with how law is taught, so that students exiting law school have a solid skill base.

Instead of thinking of them as bolt on classes, law schools should incorporate the methodologies in every class. This would require both training faculty and re-structuring classes, of course. The benefit of building the methodologies into the classes is that students are simultaneously trained on theory and practice skills. The skills become part of “just how we do things.” Some schools might embrace this approach while others might resist. Again, we can let the market decide which is the better approach.

As each year passes, it becomes more difficult to define “practicing law.” At one time, legal theory was intertwined with legal services delivery. Today, much of law is not complex legal theory, it is the prosaic implementation of long-established theory in routine settings. Efficiency, quality, and timeliness take precedence in those settings. Even when novel legal theory is involved, it quickly gives way to implementation, and again efficiency, quality, and timeliness become important.

Failing to teach law students basic tech skills and then sending them (or at least 40% of them) off to solo to smaller firms which also lack basic tech skills, perpetuates an tech ignorant profession. Now is the time to break the cycle and introduce tech skills into every law school class. Failure to do so just hastens the day when lawyers will find themselves obsolete.

Ken is an author on innovation, leadership, and on the future of people, processes, and technology in the legal industry. He also is an adjunct professor and Research Fellow at Michigan State University’s College of Law. You can follow him on Twitter, connect with him on LinkedIn, and follow him on Facebook.



Ken Grady
The Algorithmic Society

Writing & innovating at the intersection of people, processes, & tech. @LeanLawStrategy; https://medium.com/the-algorithmic-society.