Musings on AI and the Future of Law; Part 2

In part one of these musings I discussed the state, and limits, of AI in the legal world. In this article, the fun begins. What will the medium term implications of this powerful and developing technology have on the practice and the business of legal services? And what may the long-term future hold?

Transformation of the Practice of Law: 2016-2025

Looking at the next ten years, it seems highly likely that AI will become a more commonly adopted and utilized tool used by practitioners in a wide variety of settings. But AI is still nothing like a fully functioning human being. As such, I believe the most important developments in AI and the law over the next ten years will be the dissemination and adoption of the technology, in more-or-less its current forms, to a greater number of law firms, corporate law departments, geographies, and jurisdictions. This will happen along the lines of other technological advances in the practice. There will be early adopters (see part one) as well as naysayers. There will judicial opinions that slowly embrace the technology and legitimize its use in practice, similar to the adoption of predictive code in US e-discovery. There will regulatory and professional responsibility changes, similar to what the ABA has already done, to both further require lawyers to understand the technology as well as ring-fence and protect entrenched business interests of current lawyers.

[Sidebar: This does not mean to say that Kira, RAVN, ROSS, NexLP, or any of the other AI solutions on the market today (whether mentioned or regrettably not covered in part one) will simply stagnate. But given the trends for adoption of technology in law historically, as well as the professions’ track record with change and innovation, I believe that the next decade will be primarily marked by the integration of AI-powered tools into the lawyer’s everyday toolkit and practice. There will be technological breakthroughs as well, many of which will be followed and perhaps even adopted by the profession. However, I believe the pace of change will remain such that those breakthroughs will not be the primary drivers of change in the medium term.]

AI is simply not ready to replace human lawyers. Human beings are still better at negotiating. Human beings still have relationships. Human beings still have judgment and can incorporate moral, ethical, reputational and political risks into their judgment alongside legal frameworks. Even if AI impacts the understanding of the legal framework, humans will not be cut out from the equation. Human beings are still better at understanding the motivations of other human beings and crafting solutions that speak to the underlying desires and needs of counterparties and stakeholders. So long as these statements hold true, there will be a need for flesh-and-blood lawyers. At the present, AI is rational. As we can see from the ascent of Donald Trump in the US, to the Brexit vote, humans are complex creatures that may be rationalized, but are not purely rational in their actions. For now, the profession seems fundamentally safe – though there are contrary opinions out there.

I also predict that AI will fundamentally change the business of law in the next ten years. The current law firm model in the US, UK, Europe and Australia is based on high associate leverage – there are a lot of humans that expend a lot of (billable) time doing repetitive and time-intensive tasks. These tasks may include document review, contract review, due diligence and similar activities. Understanding the documents and the facts and legal terms contained within them is a critical component of a lawyer’s job, but the actual process of reviewing tens of thousands of documents – or more – strains the stamina and concentration of any mere mortal. Thus, while important, these tasks tend to burden junior lawyers in that they are both high-volume and low-discretion. The amount of time spent on these tasks by these leveraged associates makes partners at these law firms exceptionally wealthy. Or as I have said in the past, document review and diligence is what makes partners rich and clients frustrated. AI is, and will continue to, attack this fundamental business model and the firms that do not adapt to this new reality will fail, or will be acquired and transformed. This will not only attack the leverage model of law firms, but it will also attack the revenue generation model and will prompt new pricing strategies that attempt to capture the profitability generated by efficiency. While I have a lot of views on what firms will need to do to adapt, that will have to wait for another musing.

In most jurisdictions in Asia the impact will be far less tectonic. While any generalization must be viewed as necessarily ignoring nuance and detail, there are important distinctions to the way law firms in Asia operate. The fundamental business model of law firms is less dependent on heavy associate leverage, and the use of non-billable hour arrangements is far more widespread. Thus, while associate leverage is still a critical component of revenue generation, it is not as closely linked to massive e-discovery or due diligence tasks as the source of associate utilization. Furthermore, because there is generally more comfort with fixed-fee billing, there is also more comfort with the idea that technology that improves efficiency will not pose an existential threat to a firm’s revenue model, and may even improve the bottom line. Similar to the way the developing world simply skipped dial-up internet and went straight to broadband, I predict that law firms across Asia will integrate AI into document review, contract review, compliance work and to a lesser extent legal research seamlessly and without any fundamental transformations to their business models.

In this time frame, I predict that the impact on corporate legal departments will not be as dramatic. In general, corporations across the globe will enjoy more predictable costs and more controlled time frames for the completion of work. Law firms will likely develop, license or purchase technology that removes the most time-intensive and least “legal” of the tasks they currently perform.

In terms of legal employment, I do not expect AI to radically impact the demand-side of the labor market materially in the next 10 years. It is true that AI will streamline many of the tasks that currently occupy the billable time of junior associates. However, this was also true of the rise of (LPO) out-sourcing and near-sourcing of these same tasks to lower-cost centers. Even after non-US based LPO providers captured over US$1 billion in revenues in 2012 and achieved CAGR growth of over 29%, the labor market for new attorneys in the USA was stagnant, not contracting. This may also be the case in the integration of AI into law. Junior lawyers will be free to do both high-level tasks and will become the operators of these AI engines. More legal capacity will encourage more lawyering for particularly risk-sensitive clients. As noted below, I do believe the long-term effect will be fewer US-based lawyers, but I don’t believe we will truly see the impact of that in the near-to-medium term.

Transformation of the Practice of Law: 2025 and beyond

So now we are ten years into the future, and the world is even flatter than it is right now. AI is a regular practice tool for facilitating transactions and pulling together the relevant facts for disputes. This is as true in North America as it is in previously “developing” economies of the world. What further effects will this have on the profession?

First, I predict that legal education will fundamentally transform. Lawyers value proposition will continue to be disconnected from being able to recite the law as that information will be readily accessible to sophisticated lay-people. Lawyers’ value will be found in their ability to advocate, negotiate, make personal connections, and persuade. Law school curriculums will increasingly look like business school curriculums and will focus much more on interpersonal and leadership skills rather than on consuming vast tracts of legal rules and opinions. Of course, this prediction still assumes that advocacy, negotiations and persuasion will still be valued human attributes. If significant advancements are made in the field of AI so that machines become adept at both inductive and deductive reasoning, morality, ethics, cultural differences etc., the question no longer becomes what is the role of lawyers, but rather what is the role of humans in business – which is well beyond this article.

Another potential avenue for legal education to explore is to create lawyers that act more as engineers than as what we currently understand as lawyers. They may be more proficient as operators of tools, rather than as trained professionals highly skilled in the art of identifying risks and giving advice. The lawyer of tomorrow may need to be as skilled in emotional intelligence as they are in statistics, game theory and coding machine-based learning tools. In other words, the skills currently associated with counseling and advocacy will become more important. Simultaneously, skills as a technology engineer with the ability to manipulate and operate highly sophisticated AI tools with become an indispensable part of the lawyers’ toolkit.

Second, I predict that law firms will start selling technology, in the form of SaaS, to their clients for understanding basic rules. Law firms will have technology support arms staffed with sales and support professionals. Law firms will also employ full-time professionals to maintain and update the substantive accuracy of these services. Essentially, I envision a massive expansion of tools such as MarginMatrix that use even more sophisticated technology and cover a far larger portion of the corpus of laws as one of the key drivers of law firm revenue.

Another interesting prediction made along these lines is the selling of machine learning as a service. While that idea is fascinating, it poses yet another way that lawyers could identify new offerings to provide to clients, now entirely removed from the practice of law. This would transform lawyers into both engineers and product specialists providing after-sale support and customization of the various legal tools they sell to clients. I have no idea if this particular prediction will come to pass, but it a provocative future to contemplate.

Third, I predict that the largest corporate legal departments will move document analysis, both of internal information as well as review of diligence materials from outside parties, to their in-house resources, supported by AI tools. They will do this by purchasing/licensing the same AI technology that is being developed by and sold to law firms (thus suggesting another revenue source for law firms that would compete with traditional software shops). This trend will parallel the existing trend to move e-discovery in-house in the US. As this technology becomes cheaper, along with better cloud infrastructure, the size of corporation that will purchase these services will dramatically decrease as vendors move down the value chain seeking to expand their market. Alternatively, the cost of services regarding document analysis maybe come sufficiently low that rather than moving these resources in-house, specialized departments of law firms plus other technology vendors will provide these services as a vendor leveraging large economies of scale.

Fourth, I expect that the number of lawyers will increase globally, but not evenly. I predict that the disruption to the leverage model in the US will lead to fewer lawyers there. The UK and Australia will not be impacted as severely because of earlier adoption of technology in their practice and more sophisticated understanding of their cost basis for producing revenue. AI-based tools will be both a standard aspect of practice within these firms and will be a SaaS offering that will be sold by firms to clients for specific tasks as well as on a broader license basis.

I assume that in the future, the global nature of trade and markets will continue to increase, along with the rapid adoption of technology in previously developing economies and infrastructure development. When taken all together, this will increase the amount of cross-border finance, transaction, and litigation into regions that have thus far not been fertile grounds for this type of legal work. If the development trends continue, along with the current dominance of Western financial institutions as a source of capital for development, it will lead to an explosion of demand for lawyers in Asia and Africa. Cross-border negotiating skills will be highly valued as well as the ability to quickly synthesize complex and contradictory regulatory regimes. US-trained and qualified lawyers will be in high demand (assuming legal education adjusts to provide them with the skills needed for this marketplace) but they will no longer primarily be employed within the US or for US-headquartered law firms and corporations. This also assumes that US-trained lawyers will continue to benefit from the importance of the US as an economic power and center of finance. With time, it is easily conceivable that another economically significant country could produce world-class lawyers that compete or even overtake US-trained lawyers in attractiveness or skill-sets. With AI as the background for what was once specialized knowledge, the talent market for lawyers will stop being regional and increasingly become global. This will also be enhanced by the further liberalization of lawyer-regulations around the world.

But, like I said in part one, I am no Nostradamus. Since all these predictions are inherently speculative, I invite any interested reader to elaborate on their own thoughts on how the deepening relationship between AI and legal services will evolve.