AI and the Lawyer Overload

On the overproduction of lawyers and how AI can help solve the problem

Jean-Jacques Rousseau wrote that the Spaniards, during their conquest of the Americas, showed only one display of humanity. They did not take lawyers with them. They banned them from the colonies, and the French thinker regarded this decision as an act by which they redeemed all the crimes committed against the Indians.

More than two hundred and forty years later, Vice President Dan Quayle asked if the United States, making up 5 per cent of the world’s population, must have 70 per cent of all its lawyers. The numbers he gave were wrong. Nevertheless, this marked the first instance in post-war American history, or perhaps in the history of the post-war West, where the overproduction of lawyers became a topic of public debate.

Surprisingly, it wasn’t until the 1980s that economists began to explore the impact of lawyers on economic growth. Stephen J. Magee first raised the question in 1984, inspired by the work of Mancur Olson. The growing interest in this matter was likely triggered by the fact the number of lawyers doubled from the 1970s to the 1990s (and quadrupled in Washington alone).

Quayle’s statement stemmed from the debate sparked by Magee’s theses presented at the White House. The latter argued, based on studies comparing the number of lawyers in various global economies to their growth rates, that achieving optimal number of lawyers in the U.S. – one that does not depress productivity – would require reducing their pool by 40 per cent. Some highlighted weaknesses of these claims, such as the validity of comparing countries where education systems vary significantly. Setting aside quarrels over definitions, even critics admitted that so far no serious research has been done on the question: what price is the United States paying for such a large number of lawyers?

In 2023, two economists sought to avoid the charges made against Magee and limited themselves to America alone. The conclusion of their study, which examined data from 2005 to 2018 across 50 states, indicates that the number of lawyers was negatively correlated with the growth rate.

The sum effect of the overload of lawyers and their growing influence isn’t easy to calculate, as it is difficult to figure the exact losses that result from rent-seeking, talent misallocation and proliferating lobbies.

Keep in mind, however, that talent misallocation was, according to historian David Landes, the main reason why the Industrial Revolution began in England rather than in France.

Critics portray lawyers as the main agents of rent-seeking tendencies in society. Their role amounts largely to the transfer of wealth; they do not contribute to its creation, acting as a seedbed for what Jonathan Rauch called the ‘parasite economy.’ Every regulatory battle, every new lawsuit, every struggle for redistribution results in profit for them. The transfers they enforce, along with subsidies and court awards, are quantifiable. In contrast, the magnitude of wasted material and human resources they entail, not to mention systemic inefficiencies they induce, is difficult to measure.

One can only try to imagine the extent of these losses to the economy and society. Surveys like the one conducted in 2003 by the Reinsurance Association of America (as cited here) can help: according to the poll, 87 per cent of manufacturers will defend themselves at least once in a product liability lawsuit. The high costs associated with such a lawsuit are bound to have a negative impact on U.S. manufacturing output: fear of a possible lawsuit caused 39 per cent of manufacturers to abandon new product lines, and 25 per cent to end product research. Additionally, there’s little doubt that maintaining large legal staffs and refraining from entering business due to fear of lawsuits negatively impacts innovation and growth.

But do other activities of lawyers compensate for the fact that many talented people choose predatory legal activity over becoming engineers, that instead of building in the world of atoms, they create excessive litigation and inefficient laws?

Rule of law certainly has a positive effect on the economic environment. However, it would be difficult to argue that an increase in the number of lawyers is reflected in greater rule of law.

Perhaps the beneficial influence of lawyers is to be found in the realm of politics. For a long time, the legal profession in the U.S. – unlike in France, where lawyers mostly supported revolutions – has served as a bulwark of conservatism, upholding the foundations of the republic amidst changing ideological attitudes. Since the 1960s, however, it is no longer the case. As John O. McGinnis notes, both the American Bar Association and law schools are monolithically left-wing, often espousing outlandish ideas like prison abolition. Moreover, their interest lies not in preserving a transparent and stable legal order, but rather in “dynamic forms of legal transformation and the uncertainty they bring. Far from supporting a sound, established social order, they are likely to seek to undermine it,” since by doing so they increase their own income from rising compliance costs and transfer payments. The remarks from Dickens’s Bleak House remain accurate: “the one great principle of law is to make business for itself. There is no other principle distinctly, certainly, and consistently maintained through all its narrow turnings.”

***

To claim that law is a line of code of the Republic would be too simplistic. The reality is more complex. This does not mean, however, that AI will not address the problem of overproduction of lawyers.

Critics are calling our attention to first problems with the use of AI in legal practice. In New York, two lawyers cited cases that turned out to be hallucinations. Some judges now mandate that documents submitted to them have a certificate that they were not generated by LLMs, or if they were, that they have been thoroughly checked.

It seems reasonable to believe that sooner or later, AI used in legal practices will, if not eliminate, then significantly reduce the occurrence of hallucinations. Importantly, AI seems to be ideally suited for digging through large databases in a process of so-called discovery, which until now has been particularly labor intensive. Sifting through these documents has been the preoccupation of junior litigation teams – AI will reduce their numbers and make the process itself both faster and easier. A similar development will happen with the search for precedents and analogies. Fatigue and inattention will no longer be factors, and search engines will eventually suggest the cases most relevant to the one under consideration.

AI may never be flawless, but it simply needs to be better than most attorneys. As McGinnis writes, AI is not overconfident and guided by emotions, so it will be better at assessing whether it makes sense to go to court or not. The number of cases will potentially decrease and they will be settled faster.

AI will also revolutionize another aspect of legal services, that is forms. In the future, when we want to draw up a will or set up a trust, we will not have to go to a lawyer. Instead AI will generate these documents, contributing to the democratization of legal services.

The development of AI will also push the best lawyers to become even better. The new technology will favor those who constitute its ‘scarce complement,’ superstar performers whose collaboration with AI will allow them to significantly reduce the number of associates they need.

Some point out that LLMs, due to the limitation of their datasets, will stall the development of the law and won’t invent novel arguments. But does the average lawyer often come up with new arguments? It seems that creativity, where it is truly necessary, is the work of a capable minority, whose skills will only be augmented by technology.

Except for the most gifted, AI will not displace, for the time being, the category of narrowly specialized lawyers, as it will need time to absorb new data. Perhaps in the near future, we will see a dwindling class of lawyers racing to create new regulations just to maintain an edge over AI.

AI is already opening up new opportunities for non-legal talent. It is probable that many of these individuals will be engineers, who will create legal value through technological innovation. Here, however, we are already seeing resistance from the industry itself. An analysis of job postings shows that there is a clear difference between US law firms and their UK counterparts: the former want to hire lawyers with digital skills, while the latter are open to people with digital skills even without a legal training. What is perhaps even more striking is that lawyer jobs requiring digital skills were associated with better pay in the UK than those not requiring them, while in the US it was the other way around.

Some speculate that AI will shift the law more toward ‘legal formalism,’ that is, the perspective holding that every case has a correct or incorrect answer. The opposing view, legal realism, contends that there is no correct answer; instead, it must be determined based on the political values prevalent at the time. It seems, however, that if, in the current era of hyperpartisanship, AI were to align the US legal system more toward formalism and further away from political debates, this would not be such a bad outcome.

***

By squeezing out average lawyers who draw up contracts or handle simple cases, and empowering superstar performers, AI will exacerbate already high income inequality among the profession (while simultaneously decreasing inequality in consumption).

As Peter Turchin rightly observes, degrees make good proxies for interelite competition. The traditional entrance ticket to the elite is an MBA or a JD. However, considering that the income distribution among newly minted lawyers is bimodal, and taking into account student debt as well as the potential for technological progress to aggravate inequalities even further, we must conclude that most aspirants aiming to join the elite will likely never achieve this goal.

According to the author of The Age of Discord, it is these failed aspirants to the elite who will present a serious threat to America’s sociopolitical stability. Keep in mind that Robespierre, Lenin and Castro were all lawyers.

Turchin points to Goldman Sachs estimates that 44 per cent of legal jobs will disappear because of technology (that’s more than Magee demanded to revive American growth). Smart, desperate to get ahead, with their path to upward mobility cut off and “nothing to lose but their crushing student loans,” they could trigger a civil unrest or worse.

If Martin Gurri is right, then modern democracies are full of revolutionary impulses, yet it is futile to expect that they will ignite real revolutions. However, this threat to America’s sociopolitical stability should not be underestimated. In view of the numerous benefits that will come from AI’s revolutionizing lawyer jobs – reducing the parasite economy, improving talent allocation and democratizing the consumption of legal services – both populists and dynamists should join forces to mitigate the upheaval that Turchin predicts. As the late populist Kevin Phillips wrote in Arrogant Capital, “reversing the situation faces a slight technical hitch: the lawyers in the White House, Congress, the courts, and the state legislatures are the people who make the laws.” Perhaps the renewal of the American Republic depends on those who will defend technology from both entrenched elites and those who dream of joining them.