In this week’s Law Firm Disrupted we re-visit last week’s topic: What lawyers will use artificial intelligence to go from grunt work to great work? This time, a certifiable expert weighs in.
I’m Roy Strom, and here’s where to reach me if you are also an expert or just want to share your thoughts: email@example.com.
➤➤ Want to receive The Law Firm Disrupted as an email. Sign up here.
Grunt to Great Part 2: A Real Lawyer Weighs In
It’s noon in Chicago, and I just set a one-hour timer. That’s how long I’m giving myself to write this column. Consider it a test of how quickly I’ll become obsolete.
Last week I wrote a column scrutinizing an increasingly common refrain from AI companies and others that their software will not replace lawyers’ jobs; instead, it will just let them focus on higher-value work. I asked a series of questions about how every lawyer would move from “grunt work to great work.” Like, who’s going to pay a second-year associate to think like a partner? And at what rate?
I got a few answers.
Here is one from a senior legal operations professional inside a Fortune 100 company who has been piloting various AI platforms: “There is a tsunami coming to the associate ranks of the profession. And in my opinion, it is going to up-end the business model of many firms. To your point, there will not be enough ‘high end’ work to continue filling the proverbial plates of these young lawyers. No one is going to stick a first- or second-year on the expert’s deposition simply because she has no more contract review to do.”
And here is one from a law firm consultant: “Once AI gains a real foothold, I’m pretty sure there’s going to [be] a rough period for lawyers—at least in the short term. And law school admissions are back on the rise from what I’ve read.”
While I’m sure those lawyers are well-informed, I would venture to guess they have not written a 66-page article on the topic that is poised to be published in the Yale Journal of Law and Technology. That is what Alvin Lindsay, a litigation partner at Hogan Lovells in Miami, has done as co-author of a paper titled, “Lola v. Skadden and the Automation of the Legal Profession.”
For brevity’s sake, let me say right here that the article is one of the most thoughtful and well-researched I’ve read on the topic of what automation will look like in the legal sector.
Lindsay and his co-authors, Michael Simon, owner of legal technology consulting firm Seventh Samurai, and Hogan Lovells associates Lolita Sosa and Paige Comparato, have laid out a well-reasoned argument that artificial intelligence will indeed take over many of the tasks that lower-level lawyers today use to fill their time sheets.
“Many pundits and providers claim that the role of computers is merely to augment human work. Still, augmentation only goes so far until the human element becomes unnecessary,” write Lindsay, Simon and the two Hogan Lovells associates.
The authors begin with a look at the legal fight that David Lola, a contract attorney, once waged against Skadden, Arps, Slate, Meagher & Flom. Lola argued that the document review work he did as a contract lawyer was not the “practice of law” as defined by a federal statute that prevents lawyers from being paid overtime.
The paper calls it a “watershed moment” when the U.S. Court of Appeals for the Second Circuit agreed with Lola. The appellate court found, in effect, that work that can be done by a computer cannot be considered the practice of law.
“An individual who . . . undertakes tasks that could otherwise be performed entirely by a machine cannot be said to engage in the practice of law,” the opinion read.
Down the road, Lindsay and his co-writers argue that this ruling will neuter bar associations, depriving them of their ability to protect the practice of law.
“If a task can be done by a machine, then a person performing that task is not practicing law,” the authors write, extrapolating from the Second Circuit’s opinion. “This allows the capabilities of the machines to define what is and, more importantly, what is not the practice of law. As the capabilities of the machines improve, more and more tasks will become removed from what we can call the practice of law.”
The authors rely on a series of laws regarding the exponential increase in computing power as reason to avoid the belief that the law is somehow too complex for computers to some day understand and interact with. Moore’s Law, for instance, says the number of transistors per square inch on a circuit board doubles every year (now about every 18 months). Computing power grows exponentially.
Kryder’s Law says the amount of computer storage space is growing exponentially. And another law—do we really need to know the name, because I’m already at the half-hour mark and these Google searches are killing my time!—says that connectivity increases exponentially.
All told, the authors state this is reason to believe that advances like IBM’s Watson will sooner or later be applied to the law. They will also do a good job at taking over menial tasks. And that is something the authors urge the industry to understand.
“The coming fight to save the legal profession from ‘doomsday’ will not be an easy battle. In fact, it will be quite difficult,” the authors write. “Staving away from doom will require a shift in conversation within the legal community: Away from both the utopian fantasy that AI will improve everything about the profession for everyone, as well as the alarmist suggestions that AI will completely replace humans, towards more realistic explorations on how lawyers can and should use AI to augment their efficacy and skillsets to the betterment of themselves and society as a whole.”
Human lawyers will still be needed, the authors argue, for three main tasks. They will need to innovate; or to apply technology to their practice to provide a better and cheaper service. They will need to maintain accountability for AI systems by understanding them and, in a sense, regulating them. And, ultimately, they will need to provide wisdom and judgment.
The paper ends with a question: “For those in the field who are able to move beyond risk aversion, look at their practices critically and innovate, AI will serve as a catalyst for a more efficient and more complete practice of law. The question remains: Who among us is willing to make the painful shift in time to avoid obsolescence?”
I asked Lindsay what entity in the legal ecosystem he felt was best suited to do that: Companies like LegalZoom? Legal research companies like Thomson Reuters or LexisNexis? Traditional Big Law firms?
“If I were starting out in this world and was a reasonably intelligent person, I would want to be the person sitting in Silicon Valley who has some coding knowledge and can look around, see what people do and automate it. That is like the LegalZoom model,” Lindsay said. “Law firms like [Hogan Lovells], I’m very bullish on. Those that stay static are going to get clobbered. And those that find ways to innovate and add value to what’s available electronically are going to be the winners. I personally think IBM’s Watson has real potential. While at the same time probably hurting the legal profession in terms of the number of lawyers needed out there.”
Of course, that’s just one’s man opinion. But he’s probably spent more time thinking critically about these ideas than others. And even if that’s a difficult thing to do when the future is uncertain, it is important.
As for Skadden, the firm settled with Lola and his fellow contract lawyer plaintiffs in December 2015. A word on methodology:
I’m happy to report that I legitimately passed my timed writing test. This took me 41 minutes to write, as you’ll see by the photo above of my unedited ramblings. While it was mostly a joke, there is some seriousness behind it. Lindsay and his co-authors make a point in their article that the technology used to automate the law—in large part the ability of natural language processing to turn words into data—will also be used to automate other fields, including journalism. Of course, that has already happened to some extent with automated sports game write-ups and financial reporting. And while I don’t think robots are trying to take my job (they wouldn’t cover their plans to take our jobs!), I hope you can appreciate that I, too, have some nerves as I look into the future of my industry.
Roy’s Reading Corner
On AI and Knowledge Management: iManage announced Thursday that Chicago-based Neal, Gerber & Eisenberg was the first customer to sign up for iManage Insight, an AI search tool powered by RAVN Systems that iManage bought last year. A marriage between iManage and RAVN was by some thought to be a boon for the advancement of AI in knowledge management, as iManage warehoused law firm’s documents in one place, allowing RAVN’s AI capabilities to be more easily applied than in more siloed set-ups.
From the release: “We saw the opportunity to significantly upgrade our search platform with iManage Insight,” said Neal Gerber COO Sonia Menon. “It’s a powerful but easy-to-use application that enables our professionals to quickly surface key knowledge from many internal and external repositories. We’ve already started using one of Insight’s key features—Insight Knowledge Graph—to identify hidden experts and expertise throughout the firm. We are excited to take this step and we see it as a foundation for advancing knowledge management for the firm.”
On AI Judges: This wild article claims that China is already implementing AI to augment judges in their court system. The story, published on a site that appears to be part of the country’s official newspaper chain, states that iFlytek, a “leading artificial intelligence company,” can apply an AI “smart trial system” to cases of murder, theft, telecommunications fraud and illegal fundraising. Liu Qingfeng, chairman of iFlytek, said it will be used in 79 types of cases by the end of the year.
From the story: “According to Liu, who is also a deputy to the 13th National People’s Congress, the AI system can automatically judge whether the evidences are contradictory or complete enough to support a sentence, as well as which laws and regulations can be used, how previous similar cases are tried and suggest an appropriate sentence for reference.”
And thanks to Mark Greene’s ‘Today in Artificial Intelligence’ for the link to this last story.