AI cyber law scales of justiceSteven Hawking famously said, “The rise of powerful AI will be either the best or the worst thing ever to happen to humanity.” Both the practice of law and the business of law will change with the rise of artificial intelligence (AI), and law firms will need to adapt to those changes.

AI in the Legal Industry

AI has been making inroads in the legal industry for years, with AI tools becoming more effective at handling an ever-expanding set of tasks. Litigators have been using technology to help them review large sets of documents for over a decade, and AI technologies have begun to assist with other tasks as well, including analyzing briefs and memoranda to find additional cases on relevant issues and scanning key documents to help tailor legal research. Companies have even designed tools to predict the outcomes of cases. On the transactional side, some companies have begun to use algorithms to automate the review and management of contracts and to aid in the drafting of some types of contracts.

AI has been shown to produce high-quality results, performing equal to or better than attorneys in many areas. Studies have suggested that technology assisted reviews of documents can identify documents at least as accurately as a team of human reviewers. Maura R. Grossman and Gordon V. Cormack, Technology-Assisted Review in E-Discovery Can Be More Effective and More Efficient Than Exhaustive Manual Review, 17 Rich. J.L. & Tech 11 (2011). AI recently outperformed lawyers in predicting the outcome of certain lawsuits involving financial products, AI Beats Human Lawyers in CaseCrunch Prediction Showdown, Artificial Lawyer (Oct. 28, 2017), and in identifying the flaws in non-disclosure agreements. Jessica Stillman, An A.I. Just Outperformed 20 Top Lawyers (and the Lawyers Were Happy), Inc. (Nov. 9, 2018). Additionally, AI systems complete tasks in a small fraction of the time.

Given the potential benefits, it is no surprise that legal departments and law firms have been increasingly looking to incorporate AI into their operations. Bloomberg Law reported that the teams responsible for IBM’s Watson AI system have been engaging in informal discussions with prominent law firms, and that IBM is planning to announce its larger-scale entry into the legal market. Sam Skolnik, IBM Watson in Quiet Talks with Law Firms to Expand AI Offerings, Bloomberg Law (Jan. 31, 2019). A group of twelve elite law firms also joined a consortium to support Reynen Court, a technology company that plans to make it easier for law firms to utilize new technologies, including AI.

Impact of AI on Law Firms

Broadly adopting AI across the legal industry is likely to lead to a restructuring of the current law firm model by decreasing the number of lawyers needed to accomplish the same tasks. McKinsey recently estimated that 22% of a lawyer’s job and 35% of a law clerk’s job could be automated. McKinsey & Co., Jobs Lost, Jobs Gained: Workforce Transitions in a Time of Automation (December 2017). Many articles written on the topic of automation in the legal sector have welcomed the change. These authors predict that lawyers will be able to focus their time on the more engaging and intellectually stimulating aspects of lawyering, such as developing creative ways to present a case, cleverly structuring transactions, and engaging in oral advocacy. E.g., Gary Marchant and Josh Covey, Robo-Lawyers: Your New Best Friend or Your Worst Nightmare, American Bar Association (Fall 2018).

However, if lawyers can suddenly provide legal services 22-35% more efficiently, that means a smaller number of lawyers will be able to handle a larger number of clients. There is no evidence that the legal industry will see a corresponding increase in the volume of legal services sought, so competition for legal jobs will likely intensify as AI is more widely adopted. The current imbalance between the number of law school graduates each year and entry-level law positions is therefore likely to widen. But the structural effects of AI are likely to go far beyond the number of lawyers needed to do legal work. What will AI change? In a word: everything.

Restructuring Law Firms in the Age of Automation

“It is impossible to predict the time and progress of revolution. It is governed by its own more or less mysterious laws.” Vladimir Lenin.

Notwithstanding Lenin’s warning, we have six very tentative thoughts on how AI might affect the legal profession:

(1) Lower billings. Fewer attorneys will be able to do the same amount of work. Unless there is untapped demand for high-end legal services, which seems unlikely, the result will be fewer hours billed by law firms.

(2) Junior lawyers disproportionately affected. The reduced billings will likely result in fewer legal jobs, and this effect will be concentrated at the junior level. Higher-level tasks will still need to be performed, but fewer lawyers will be needed to support the senior lawyers performing those tasks.

(3) Reduced leverage. The reduced need for junior associates will accelerate the modern trend of reducing leverage at law firms. Since 2007, there is evidence that the model of leverage law firms have traditionally employed is facing pressure from technology and in-house legal teams. James Park, Law Firms in Los Angeles After the Financial Crisis, Lowell Milken Institute Policy Report (Mar. 2019). As AI increases the efficiency with which junior associates can complete their tasks, even fewer junior lawyers will be needed.

(4) Ending billable hours. Increased reliance on AI may hasten the demise of the billable hour. In recent history, clients have been leading the charge for alternative fee arrangements, but with fewer hours required to perform legal services, firms may begin to feel that billable hours do not adequately compensate them for the value of the work they do. As a result, firms may join clients in pushing for alternative fee arrangements.

(5) Effects vary by practice area. Different areas of law may be affected differently by AI. In litigation, AI may play an increasing role in reviewing documents and in allowing clients to more accurately assess legal outcomes, but the outcome of litigation will presumably still be decided by human judges and juries. Even in the areas of law where technology has been implemented to assist judges in making decisions, the judge still retains the full authority and responsibility to issue the decision, only using the algorithm’s report as one data point among many. Ephrat Livni, In the US, Some Criminal Court Judges Now Use Algorithms to Guide Decisions on Bail, Quartz (Feb. 28, 2017). This will continue to leave a key role for human advocates in litigation, albeit a role performed with fewer junior associates. Moreover, there may be certain types of litigation which happen so infrequently that AI may not have not have enough available data to accurately assess the outcome. But routine transactional work may well be done in-house thanks to AI, and the definition of “routine” work may become broader as AI improves.

(6) Changes in training lawyers. Perhaps most troubling, the traditional model of training attorneys is very much at risk. Attorneys have traditionally been trained by doing routine work while watching more senior attorneys. That model has already been affected by e-discovery, and with AI allowing machines to do more of the routine work, it is at best unclear whether there will be enough billable work for the (diminished) pool of junior lawyers to perform, and likewise unclear how many will receive proper training.

Because lawyers and law firms are vital to the functioning of the modern economy, it is safe to assume that some new system of training and a new balance between clients and lawyers will arise. And it is possible that one effect of AI will be to make the (presumably smaller) firms of the future look and feel a lot like the firms of decades ago, with less leverage, less emphasis on billable hours at the junior level, and more focus on training for training’s sake. But it is also safe to assume that every guess about the details of that future—including ours—will be wrong.

Lawrence T. Gresser is the managing partner of Cohen & Gresser, which he co-founded in 2002. Marvin J. Lowenthal is an associate at the firm.