This is Part II of a two-part roundtable discussion. Click here to read Part I!
In this roundtable discussion, seven prominent leaders in law and technology explore the state of AI adoption in the legal profession and what the future holds.
Nicholas Gaffney (NG) is the founder of Zumado Public Relations in San Francisco, CA and is a member of the Law Practice Today Editorial Board. Contact him at email@example.com or on Twitter @nickgaffney.
|Mauricio Uribe (MU), a partner at Knobbe Martens, is experienced in all aspects of intellectual property law, with an emphasis in the electrical engineering and computer software fields. He provides counsel on patentability, due diligence and infringement mitigation matters and comprehensive intellectual property programs.|
|Kim Massana (KM) is CEO at Neota Logic, a leading AI automation platform for professional services. Before Neota Logic, he worked for Thomson Reuters and led a number of software companies.|
|Nicolas Economou (NE) is the CEO of H5, a firm that pioneered the application of scientific methods to electronic discovery. He chairs the Law Committees of the IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems and the Global Governance of AI Roundtable.|
|Bobbi Basile (BB) serves as a managing director in the Legal Transformation + Innovation practice within the Advisory Business at HBR Consulting, and has over 25 years of experience delivering strategic, operations and technology services to Fortune 500 law departments and law firms.|
|Ben Levi (BL) is co-founder and COO of InCloudCounsel, a legal technology company that automates and enhances high-volume legal processes for large companies. Prior to InCloudCounsel, Ben practiced corporate law at Kirkland & Ellis.|
|Jon O’Connell (JO) is a corporate attorney and founding member of Atrium’s legal team. His practice focuses on representing emerging companies and venture capital investors. As a partner at Atrium, Jon enjoys advising startup clients and working with executive management to solve complex business issues in a timely manner.|
|Jeremy Peter Green (JPG) is a trademark attorney, software developer, and founder of JPG Legal. In 2018, JPG Legal was ranked the #22 law firm in the United States based on number of federal trademark applications filed. Jeremy is currently developing an AI-powered trademark clearance search engine.|
NG: What kind of legal services do you see AI playing a key role in five years from now? Twenty years from now?
MU: Generally, AI technologies are best suited for processing larger data sets, especially when the relationship between individual data or groupings of data is not immediately clear. In situations in which data sets from which decisions are made are more finite or in which the relationship between specific data and an outcome/decision is well defined, AI-based solutions may not be the best technical solution, especially considering the cost and computing resource consumption required for AI-based solutions. For example, auto-populating a document based on fields of a form does not require an AI-based solution.
In the relatively short-term future, some types of legal services and AI-based solutions may be able to increase attorney efficiencies, such as generating initial draft or sample language for contracts once a term sheet has been negotiated or identifying inconsistent language/terms in documents. But, it would seem that these types of AI‑based solutions would function more as a tool for attorneys and would not be considered key to the execution of those types of services.
In the longer term future, the greatest potential for AI-based solutions to have a “key” role are those tasks that either traditional computing solutions or human attorneys cannot execute well (or at all). For example, with regard to due diligence or discovery, AI solutions may be able to analyze millions of pages of documents not only to identify “related documents,” but to do so by also determining the criteria that define the data relationships without human intervention. In another example, AI-based solutions may be able to access large case law data sets to identify, correlate and prioritize legal precedent on a scale or within a time frame that would be impossible or prohibitively expensive under traditional methods. These types of more complex, AI‑based solutions will continue to develop and evolve, and will likely be readily available to law practitioners.
NE: I am among those who believe that, in the absence of truly disruptive innovation, AI over the next two decades will progressively make certain tasks more effective and efficient, but not in really unpredictable ways: Artificial General Intelligence is, probably, a long way off. Increasingly sophisticated technologies, operated by increasingly sophisticated and scientifically trained operators, will make law firm management, M&A due diligence, discovery, data risk management, and compliance more automated, continuing what we have seen so far.
BB: In five years, we expect to see continued advancement and wider application in:
Research, document/content assimilation, risk diagnostics and analysis (e.g., IP theft detection), contract analysis and generation, outcome prediction, etc. Progress will be measured by the seamlessness of its use, with end users unaware that machine learning is being applied as is the case with consumer products (e.g., music applications).
In 20 years, who knows? We may well be looking for ways to apply organic intelligence after the algorithms start acting like petulant humans!
BL: Broadly speaking, today’s AI technology can be classified as “weak” or “narrow,” which means that its algorithms can only be taught to tackle single, specific tasks. This reality is likely to persist for the next five years as well, meaning that AI’s application to legal services will continue to be towards accomplishing narrow and specific tasks, such as finding terms in a set of documents or filling out various forms. The role of the attorney will still be vital to conducting quality legal work.
As AI improves—and potentially becomes less narrow in scope—over the course of the next couple of decades, its impact on the practice of law is also likely to expand in ways that we don’t fully understand yet. Some of those possible areas of impact include the ability to generate agreements, to mark-up and even negotiate a document, and to automatically administer and make appropriate filings. Beyond that, it has the potential to even change the way in which contracts are drafted as word documents, which perhaps may not even be the most efficient way to handle contracts in the distant future.
JO: We are working on some pretty exciting ML and AI innovations but can’t give away our product roadmap!
JPG: Trademark clearance search engines, contract review tools, and eDiscovery tools will all probably advance significantly over the next five years. In 20 years, we’ll likely have legal brief drafting software that makes the drafting of jurisdiction-specific boilerplate language and citations much less painful by accurately suggesting what language and punctuation to use when writing court filings. These advances will be good for legal consumers and the few law firms that charge flat fees or monthly/quarterly retainers, and bad for firms that use the billable hour model, as clients start expecting the use of these tools.
NG: What influence do you think AI has had on evolving law firm business models?
MU: One of the biggest opportunities for influencing or evolving law firm business models is with regard to better understanding law firm operations. Law firms often struggle with understanding relationships between various law firm activities and associated operating costs. For example, many firms struggle to conduct meaningful return on investment analysis based on individual and group business development activities other than the traditional “what billings resulted from this activity?”
In another example, law firms often struggle to understand relationships/dependencies between staffing costs (salaries and benefits) and law firm revenue (based on attorney billings). Often, such relationships are unique to specific types of legal work, regions or even individual firms such that there is simply no universal solution/equation that can effectively model the relationship for all firms.
Unfortunately, I don’t believe that AI has had a significant influence on evolving law firm business models yet. But I assume that law firms that are considered well-managed or better run will likely be leaders in this area.
KM: AI gives law firms the ability to offer new services to their clients, that they couldn’t have offered before. For example, our law firm clients offer ‘self-service’ portals where their clients can receive legal advice, in the same way as if they had asked a lawyer at their firm and receive the same answer. The advantage being is that they can be served 24/7 and receive answers in an instant, without delay. These portals often automate work that law firms couldn’t bill for, but would still allocate their time to answer. The answer the self-service portal gives them might suggest whether they need formal legal representation from their law firm, expediting the allocation of billable work. This is a game-changer in increasing efficiencies, and we’ve seen our clients win work off the back of offering these services that their competitors do not.
NE: So far, I believe that the impact has been small. The core business model of law firms has not changed, and AI has not produced major disruption yet, nor am I sure that it will in the foreseeable future. If the hourly billing model changes, it won’t be in any substantial part because of AI, any more than it changed because of the appearance of other automation technologies like MS Word. If anything, the increasing use of AI in corporate settings and the related compliance requirements create an increased need for new types of advisory legal services.
BB: AI hasn’t changed the business model but it is being applied selectively to improve the management of the business and/or enhance or expand service offerings to clients.
JO: AI (and more generally, legal technology) has been slow to make traction due to the partner-led business model of traditional firms. Every dollar that is invested in technology, operational efficiency, or process is taken out of the partner’s pocket and therefore their share of profits at the end of the fiscal year. Based on this business model, traditional firms are not incentivized to be efficient. Instead, they bill more.
The traditional model is not client-centric, and firms like Atrium are challenging this model. As a partner at Atrium, I’m driven by providing our clients with a better experience through legal expertise, technology, and process improvements.
Similarly, our team has OKRs (objectives and key results) in place each quarter. Instead of measuring success solely based on billable hours, we set shared goals for the team and look to achieve these by providing better service, more efficient workflows, and more transparent relationships with our clients. Our business model is fundamentally different since we bill on a subscription or project basis. Since the billable hour is not Atrium’s core unit of measure, our legal teams are incentivized to use our technology to get more efficient.
JPG: Flat fee and subscription-based (monthly/quarterly retainer) billing models are making up a larger portion of the industry as attorneys come to realize they can exploit technology to charge less than their competitors for the same services, while spending less time on the same tasks, resulting in a higher effective pay rate per hour worked. Rather than spending two hours on a legal task at $300 an hour for a client and charging the client $600, an attorney using advanced AI software can spend one hour performing the same task and charge the client a $450 flat fee, benefiting both the client and the attorney. This kind of mutual benefit from harnessing AI is only achieved if the attorney charges for the services performed (flat fee) rather than the time spent performing the services (hourly billing).
NG: Is it an overstatement to say “AI will revolutionize the legal profession,” as some experts have suggested?
MU: Attorneys have been faced with the opportunity and challenge of incorporating technological advances in the legal practice for many years. It is difficult to imagine that any technology, including AI-based solutions, would radically or fundamentally change the practice of law over a relatively short period of time (e.g., within one or two years). However, the continued advancement of AI and AI‑based solutions will modify the practice of law and client expectations regarding representation/services. Some roles and activities will likely go away. Other roles or activities will be modified in view of AI. New roles and activities will emerge (e.g., data managers). Measured over a longer period of time (e.g., five to 10 years), such changes may, in fact, be viewed as the revolution of law by AI.
NE: I believe that AI will revolutionize the legal profession, but in a different way than is typically thought. AI for most important real-world legal applications will require more, not less, scientific expertise than is currently applied. In nearly every area of human endeavor, new technologies (like AI) and domains (like data science) have resulted in the emergence of independent professions and industries, taking over from early adopters. This will be true of AI in the service of important legal functions. AI will revolutionize the law, but mostly from the outside rather than the inside. In general, lawyers will never become great data scientists any more than they aspire to be great forensic accountants. But lawyers will increasingly rely on emerging professions in AI and data science to help them enhance the efficiency of their practices, win cases, develop data risk-management practices, and help clients meet their regulatory obligations.
BB: Yes. AI is a raw technology. It requires more than technology to revolutionize anything. Sound leadership, good lawyering, and progressively-minded clients are necessary to foster the transformation of the legal industry.
BL: It depends on what we mean by “revolutionize.” AI will certainly continue to change the profession. Tasks that took hours might take nanoseconds. Data that was inaccessible may become immediately available and actionable. Law firms might have engineering departments and product teams. I view these potential changes as an opportunity for lawyers. They will be able to tap into AI-enabled legal tech solutions that will allow them to complete more work at a higher degree of accuracy, freeing up bandwidth to focus on different and more complex types of work that can create substantial value for their companies or their clients. Further, part of their role will be working with technology to train it on data sets and nuances in the law. Deep legal expertise is required to create technology that operates in the legal arena. That knowledge currently resides in humans, and that won’t change in the foreseeable future. So, rather than a revolutionary change to the profession, I suspect we will see a reimagining of what it means to be a lawyer, and what it means to work at a law firm or as an in-house counsel.
JO: This is not an overstatement. The value that AI can provide to the legal industry is just getting started. At Atrium, we believe that the industry is overdue for a change. It is time firms provide clients with the service they expect, and redefine the attorney-client relationship as a strategic advisor to your business.
JPG: I wouldn’t quite say AI will revolutionize the legal profession. It will primarily shrink it a bit more, as technology has been doing for decades. The legal industry is harder to revolutionize with technology than other professions because certain tasks will still require a person with an American bar license. So you may have one attorney performing an amount of work that would have required two attorneys in the past, creating an even more dire job market for attorneys than before, but there’s no real danger of U.S. attorneys being completely replaced by AI, the way there is for long-haul truck drivers.
NG: What does the increased adoption of AI mean for law students and young lawyers today? (For instance, will the fear of fewer jobs come to fruition? Should they be focusing on acquiring skills lawyers have traditionally not needed?)
KM: At Neota we actually run our own education program, partnering with universities and law firms. We teach students how to use our AI automation platform to build real-world solutions for not-for-profit organizations.
We feel it’s important to educate students on what AI really is so they are clear that it’s not going to take their jobs. We’ve actually found it’s creating more jobs. Many law firms reach out to us with new “legal tech” and “innovation” job descriptions and ask us to forward these to our alumni students. Law firms are crying out for applicants with an interest and understanding of emerging technologies, as well as a legal background.
NE: I think in the short term, the emergence of AI can make the work of young associates more interesting. Interacting with AI systems and, more importantly, the data scientists that increasingly operate them is far more interesting and educational than doing document review. I believe that young lawyers would benefit from understanding what skills and competencies are needed to be successful in managing groups, vendors, and service providers that use AI to deliver important legal services. But I do not believe that young lawyers can, as a rule, aspire to learn the complex data science and statistics skills needed to be—in addition to a great litigator or corporate lawyer—a great AI operator. The lawyer of the future will be effective in the use of AI if they have the knowledge a generalist manager has, not by trying to learn statistics in law school.
BB: The impact of AI and other advanced technologies is two-fold: 1) Businesses are more complex than ever, and attorneys will be expected to understand the nuances of how technology supports the business and the associated interpretation of laws that may not have kept pace; and 2) Attorneys will need to be adaptable to technologies that support the practice of law, and willing to continuously improve how law is practice and the client experience. The days of rigidity are fading as every aspect of our lives is supported by technology and fluid upgrades. The practice of law needs to align with a mindset of constant transformation to enhance the client experience and personal and professional fulfillment.
BL: I think we need to redefine what it means to work in the legal field. As we have seen with automation in other industries, there may be fewer traditional legal jobs because some of the traditional work can be done more efficiently and with fewer bodies. However, legal tech is itself a growing sector of the legal industry, and it, too, requires workers. I don’t think it’s clear that there will be fewer jobs in the legal field at large. Rather, efficiencies in an industry can lead to more jobs. As costs come down due to efficiencies gained, more people may take advantage of legal services, meaning more providers and different kinds of providers will be needed, potentially leading to job growth.
The legal profession, like many others, will continue to place a high premium on attracting thoughtful and intelligent professionals. Young lawyers should not fear that they won’t be able to find work in the industry. They should, however, think more broadly about what it means to work in the legal industry and what kinds of skills—beyond legal reasoning and writing—are required to be successful in that industry. As all industries move towards the adoption of even more and more powerful technology, young lawyers should invest in technology skills in addition to their traditional legal education.
JPG: AI doesn’t affect the job prospects of law school graduates as much as the fact that we have too many accredited law schools graduating too many new lawyers, and there are too few legal jobs for them. The legal industry is already abysmal. Prospective law students should be making sure they go to the highest ranked schools they can, preferably with full scholarships. Most of them should avoid law school altogether. Marketing, software development, and web design skills might help some law graduates at the margins if they want to hang their own shingle, but no skills will change the reality that there aren’t enough jobs for the law graduates that schools are churning out, and the salaries and benefits of about half of those jobs are declining, while law school gets more expensive. AI will hasten that, but it won’t fundamentally change the state of the job market.
NG: In what ways has AI made human skill more important than ever?
KM: In the case of the legal industry, I think it really highlights that lawyers are really needed to do what they are trained to do; offering complex, high-value legal advice. Also, machines cannot replace the relationships, empathy, and trust that firms build up over time with their clients.
NE: This is an excellent and, I would imagine for some, counterintuitive question. Isn’t AI, after all, supposed to make our lives easier, as Siri does? In fact, in high-stakes legal applications, AI requires considerable human expertise in fields like computer science, computational linguistics, and statistics to effectively achieve complex data analytics tasks. This is why there is such a surge in demand for individuals combining those skills with experience in law-related applications such as discovery. But we all must remember that human intelligence—with the ability to grasp and adapt to nuance, intuit and respond to expectations and layered meaning, and comprehend the practicalities of human experience—continues to be a critical component in any AI endeavor.
BB: The nexus of problem-solving in the legal profession is understanding the problem, framing the questions, interpreting the law and applying reasoning within nuanced fact patterns, in order to develop and guide the legal strategy. Opportunities abound when viewing AI technology as an enabler to attorneys. The lines are blurring between traditional roles in the legal profession presenting an unprecedented chance to reshape careers and businesses alike.
BL: The profession will place an even higher premium than it already does on thoughtful, intelligent and creative professionals, rather than looking for people who are simply willing to put in the long hours that being a lawyer currently requires. Beyond that, lawyers who can understand the technology and work closely with engineers and product teams will also find it easier to successfully take advantage of the technology tools that are available to them or to even build their own.
NG: Are there any legal jobs you expect will be entirely replaced by AI technology?
NE: I think legal work will evolve, but I don’t foresee entire categories of legal jobs being eliminated. There are too many unmet needs, and AI itself is creating many of them, for example in regulatory compliance and data privacy and security. A good analogy is the financial services industry: the emergence of Excel and other sophisticated financial tools did not eliminate jobs in those industries. It helped them grow and offer ever-more diverse services to an ever-larger segment of the industry and citizenship.
BB: Administrative and paralegal roles along with first- and second-year associate roles will need to evolve to adapt to the new emerging paradigm that requires different responsibilities and skills.
BL: Honestly, no. I think some legal tasks will be entirely replaced by AI, but very few jobs in the legal industry (if any) only serve to perform single tasks or tasks that can be accomplished at high quality without the direction of expert human lawyers. I think it is actually far more likely that AI will lead to new kinds of legal jobs, particularly those in legal tech, than it is that AI will cause any legal job to totally disappear.
JPG: Law librarians will be replaced entirely. At first by local tech support workers, and then by a combination of foreign-based tech support workers and AI help desks. On the other hand, law firms will employ more data analysts for marketing and internal evaluation, and data analysts are sort of the younger cousins of librarians. Many legal assistants and paralegal jobs will also be eliminated, but many attorney jobs will be replaced with the legal assistant and paralegal jobs, so I don’t think any of those three primary legal positions will be altogether eliminated.