Thanks again to all my readers! Grateful you’re here. ❤️

And to all of you newbies, welcome to the club! 👋

This substack, LawDroid Manifesto, is here to keep you in the loop about the intersection of AI and the law. Please share this article with your friends and colleagues and remember to tell me what you think in the comments below.

Thanks for reading LawDroid Manifesto! Subscribe for free to receive new posts and support my work.


Let’s take advantage of this leap day, to leap into the unknown together! I want to take a look at the “AI will replace us” narrative from a different perspective. Namely, one that is grounded in history and facts, rather than an emotional knee-jerk response. What can our experience with other innovations tell us about what to expect from this intelligence explosion? What are the commonalities and the differences? And, how can we use that insight to our advantage?

If this sounds interesting at all to you, read on…

Building a Narrative: Machine Conquest

The “Machine Conquest” narrative is predicated on various reports and studies that analyze current technological capabilities and predict their potential impact on the workforce and economy. Their conclusions have in turn been sensationalized by media outlets driven by economies of attention and controversy.

One such study examined the ability of computers to carry out various job functions. Asking the question, “how susceptible are jobs to computerization?,” the authors found that nearly half of all U.S. jobs (47%) are highly likely to be automated within the next ten years or so. This study has been widely reported with much dramatic flair.

Now, if you think that’s a reference to last year’s Goldman Sachs report, your confusion is understandable. The findings sound quite similar: “The investment bank’s economists estimate that 46% of administrative positions, 44% of legal positions, and 37% of engineering jobs could be replaced by artificial intelligence.”1

But, the first study I referenced was published in 2013 by Oxford professors Carl Benedikt Frey and Michael Osborne, who used a machine-learning algorithm to assess the adverse impact of computers on jobs. (As a side note, the US economy added 18 million jobs between 2013 and 2023.2)

Historical Context: ATMs and bank tellers

Automated teller machines (ATMs) made their debut in the United States during the 1970s, taking over several tasks traditionally done by bank tellers, like dispensing cash and accepting deposits. Unlike human tellers, ATMs could perform their jobs 24/7/365, without need for vacation or sick days, and only requiring occasional maintenance and restocking. From the mid-1990s, banks began to significantly expand their ATM networks. Although there was a notable decrease in 2020, the current number of ATMs ranges from 520,000 to 540,000.

When ATMs were first introduced, the prevailing belief, which persists today, was that if a machine could automate a task previously done by humans, then these machines would take over that task entirely and eliminate the human jobs that relied on it. (The Machine Conquest narrative at work.) A New York Times article, published in 1973, opined that ATMs would replace 75% of human teller jobs.3 But that did not happen.

Instead, tellers jobs have grown slightly faster than the general labor force. We now have nearly half a million ATMs and the same number of tellers. ATMs simply have not replaced humans. And, given a shift nowadays from cash to electronic forms of payment, the number of ATMs stands to decline over time.

But, why didn’t ATMs eliminate human tellers?

ATMs reduced the number of human tellers needed at a bank branch.4 This reduction in the number of human tellers needed per branch made running them cheaper. Because branches were cheaper to operate, banks opened more branches. Because more branches were created, more human tellers were needed.

If the analysis had stopped at the first sentence (ATMs reducing the number of teller jobs), one would have concluded that the Machine Conquest thesis was correct. This is because we underestimate, or fail to recognize, the impact of second-order effects. Every action has a consequence, and each consequence has a subsequent consequence that we may fail to appreciate. In this case, ATMs increased the demand for tellers by improving the economics of running a bank branch.

ATMs freed up bank tellers to focus on higher value work. Banks refocused human tellers on relationship banking, small business customers, and high-margin financial services. In person, human relationships led to greater profits for banks and better job security for the tellers who created that value.

Another Example: Drones and Pilots

Ten years ago, the same fear-induced panic struck with the proliferation of drones. Pilots feared for their jobs. Drones would eliminate the need for pilots in everything from cargo transportation to surveillance to air combat.

But things have not played out that way. In fact, the opposite has happened: pilots are now in such high demand that a pilot shortage has become a national problem.

The U.S. Air Force now has more jobs for drones a/k/a Remotely Piloted Aircraft (RPA) than any other type of pilot position according to the head of Air Education and Training Command.5 For comparison, in 2017, there were 1,000 drone pilots versus 889 airmen piloting the C-17 transport and 803 flying F-16s. The Air Force hopes to bump up the pilot production line. “We produced 1,108 graduates from our pilot training program last year; we’re going to produce 1,200 this year,” Lt. Gen. Darryl Roberson said. “And we’re on our way up to producing 1,400 with the assets in the U.S. Air Force … and even that’s not going to be enough to meet our requirements.” Roberson said the plan is to train 1,400 airmen in subsequent years.6

Drones drove down the cost to manufacture a fighter jet. MQ-9 Reapers cost $32 million a piece as compared to the $150 million price tag for a manned F-22 Raptor. Because drones are less expensive to manufacture, the Air Force has been able to purchase more of them, and needs more pilots to operate them.

And that’s just pilots. Drones have multiplied the total number of jobs by the massive increase in support staff needed for everything from drone logistics and maintenance, to operations and intelligence gathering.

AI and Lawyers: Drawing Parallels

Are the ATM-teller and drone-pilot experiences predictive of what will happen with AI and lawyers? It’s hard to say, but there are some lessons we can derive from them.

In both, the introduction of the technology caused initial human job loss. This job loss resulted in the more cost effective delivery of financial services with ATMs and air superiority with drones. Because these services could be delivered more efficiently, the supply of bank branches and drones increased to more efficiently meet demand. This scaling up increased the total number of jobs for people, including the creation of skills and job descriptions that did not exist before.

There are some parallels.

Certainly, AI and, in particular, deployment of Generative AI will eliminate the need for as many lawyers as they have traditionally been defined. This may cause the delivery of legal services to be delivered more cost effectively. And this efficient service delivery may increase the total demand for legal services by increasing the total addressable market. But, this can only happen in the legal industry if regulations and ethical rules don’t prevent this transformation. (Bankers did not sue ATM manufacturers for the unlawful practice of banking.)

Unlike ATMs and drones, the use case of AI in the legal industry is focused on knowledge work, including analysis, drafting, research and other soft skills that have heretofore been beyond the grasp of automation. We may think of this as different because it feels that work that we do is more “in our head,” rather than out in the world and mechanical, like dispensing cash or flying surveillance. But, lawyers still produce work product, namely, pleadings, motions, demand letters and the like that needs to be filed in a court or relayed to a client.

But is AI different?

One can see the similarities and think that legal services is safe from encroachment because it will follow the same pattern as ATMs and drones and result in more jobs. Of course, it is comforting to think that may be the case. But AI makes it possible now to not only automate mundane tasks but also more abstract and strategic skills.

Once there is a “full spectrum” solution like this available, what is left for lawyers? Jordan Furlong argues that what is left is for lawyers to advocate, advise and accompany. Certainly, the human relationship lawyers have with their clients remains, to act as a guide through a stressful situation, to provide peace of mind and support.

Less certain is the need for as many lawyers as they have been traditionally defined. As market forces cause AI to climb and assimilate the legal services value chain, lawyers will have to specialize and retrain to accommodate the new paradigm. Yes, there will be a need for innovation managers, legal engineers, data scientists and legal service designers, but also still for client relationship managers and rainmakers.

Closing Thoughts

In our journey through the narrative of “Machine Conquest,” we’ve observed how technological advancements, from ATMs to drones, have reshaped the workforce in ways that were initially unforeseen. These examples serve as historical testaments to the adaptability and resilience of human employment in the face of automation. They underscore a critical insight: that the introduction of technology not only transforms jobs but often expands the ecosystem of work in unexpected directions.

As we turn our gaze towards the future, particularly with AI’s burgeoning role in the legal profession, it’s tempting to forecast doom and disruption. However, the lessons of the past urge us to adopt a more nuanced perspective. Just as ATMs and drones have not spelled the end of bank tellers or pilots but instead have evolved their roles, AI’s integration into legal work promises a similar trajectory of transformation rather than outright displacement.

The emergence of AI in the legal field necessitates a redefinition of what it means to practice law, emphasizing the uniquely human skills of empathy, ethical judgment, and creative problem-solving. In this future, lawyers will increasingly act as orchestrators of legal solutions, integrating AI’s analytical prowess with their deep understanding of human needs and societal norms.

As we navigate the complexities of AI integration, our focus should not solely rest on the jobs it might replace but on the opportunities it creates for enhancing human expertise and fostering new forms of collaboration. The future, then, is not about machines conquering the human spirit but about harnessing the symbiotic potential between human ingenuity and artificial intelligence to forge a more nuanced, efficient, and empathetic legal landscape. “Collaborative Intelligence” is our new narrative.

Now, let’s go make that happen!


By the way, if you’d like to learn more about how how AI will impact the legal profession, I’ve launched a free 5-part webinar series, “Generative AI for Lawyers: Empowering Solo and Small Firms.” You can learn more about it here.

1

Two-Thirds of Jobs Are at Risk: Goldman Sachs A.I. Study, The Observer, March 30, 2023, https://observer.com/2023/03/generative-a-i-may-replace-300-million-jobs-goldman-sachs-study/. Also see 3 Geeks and a Law Blog’s detailed analysis of what the Goldman Sachs study actually said versus what was published in media reports, July 31, 2023, https://www.geeklawblog.com/2023/07/44-of-investment-bankers-think-they-can-make-lots-of-money-off-of-attorney-insecurity-ai.html

2

Number of full-time employees in the United States from 1990 to 2023, Statistica, https://www.statista.com/statistics/192356/number-of-full-time-employees-in-the-usa-since-1990/

3

Machines – the New Bank Tellers, New York Times, December 2, 1973, Page 219, https://www.nytimes.com/1973/12/02/archives/machines-the-new-bank-tellers-response-to-automated-transactions-is.html/

4

ATMs reduced the number of bank tellers necessary at each branch, from 20 to 13 between 1988 and 2004. https://www.imf.org/external/pubs/ft/fandd/2015/03/bessen.htm

5

Drone Milestone: More RPA Jobs Than Any Other Pilot Position, Military.com, March 8, 2017, https://www.military.com/daily-news/2017/03/08/drone-milestone-more-rpa-jobs-any-other-pilot-position.html

6

Id.