Section Review

In Yuval Noah Harari’s bestselling book, 21 Lessons for the 21st Century, the author discusses the impact of technological advances, especially artificial intelligence (AI), on the future of the global job market. Harari envisions a not-too-distant future where AI advancements could cause significant job losses. The legal industry, in his assessment, would not escape this fate. Harari opines:

Good drivers, bankers, and lawyers don’t have magical intuitions about traffic, investment, or negotiation; rather, by recognizing recurring patterns, they spot and try to avoid careless pedestrians, inept borrowers, and sly crooks. It also turns out that the biochemical algorithms of the human brain are far from perfect … No wonder that even good drivers, bankers, and lawyers sometimes make stupid mistakes. This means that AI can outperform humans even in tasks that supposedly demand “intuition.”

While Harari looks several decades into the future, concerns about uncontrolled AI have been raised by other figures in recent years, including Richard Susskind and Elon Musk. While Musk’s cogency on a variety of issues is becoming increasingly uncertain, he stated in 2014 that AI posed an existential threat to humanity. 

Someone reading and hearing these things about AI generally and in specific relation to legal practice may be thinking to themselves — “if the machines are coming for us, what’s the big deal if I don’t hit my billable targets for this year or reel in that long-sought-after client?” Those thoughts, serious or otherwise, tend to arise from unfamiliarity with AI and its ubiquity in legal practice already. Not to mention the many benefits that AI technologies have had and will continue to have on legal practice. At least in the humble opinion of this scribe. 

All that being said and now that I hopefully have your attention — what exactly is AI?

AI is machine — not human — intelligence. Generally understood, the three categories of AI applications are: (1) robotics; (2) natural language processing; and (3) machine learning. 

Robotics

Robotics broadly refers to instances where a “robot” is used in place of a human to perform automated tasks. There is presently no clear robotics application in law, but we will continue to monitor this space. 

Natural language processing

Natural language processing (NLP) concerns technologies that can understand and interact with text and the spoken word. Siri and Alexa are prominent examples of NLP in the consumer context. There are limited examples of NLP already in legal practice (e.g., conversational legal “robots” or chatbots on firm websites for client review and intake). An area for future growth is adapting digital assistants for use in the legal context. Digital assistants demonstrate NLP through the translation and analysis of human speech to perform a task, whether it be providing the weather forecast, answering a question or setting a timer. Digital assistants typically deal in this lower-level consumer information.

Digital legal assistants, on the other hand, would implicate professional-grade advice. Practitioners would ask their machine counterpart for information on a case holding or statutory citation, instead of taking the time to manually look it up in a treatise or online. Professional-grade assistance presents much higher stakes, in which the user is relying on professional advice and recommendations that have more significant consequences. Therefore, these assistants can create potential issues for lawyers when considering the looming threat of a malpractice suit. 

Incorrect professional advice can have serious consequences for the practitioner and client. This is much different from consumer information, such as a person’s horoscope or the day’s forecast. For instance, if the weather forecast is wrong, and you don’t bring an umbrella, you might get a little wet. If someone relies on consumer-grade medical or legal advice, on the other hand, that may have life-altering effects. And when it comes to a legal assistant, relying on its recitation of a case holding or statutory text, instead of independent verification, likely falls short of satisfying one’s professional responsibility. The complexities of legal practice, and its many quirks, pose unique challenges for the development of an effective digital legal assistant. 

Machine learning

Machine learning involves efficient data collection and analysis, based on pattern recognition, which is then used for outcome prediction. Various machine learning applications have already established a firm foothold in legal practice. These include e-discovery/document review, con-tract review, and legal research engines. Other technologies have appeared on the stage and will likely feature more prominently moving forward: litigation finance; automated timekeeping and bill review; and legal marketing, to name a few.

Ethical considerations

AI in legal practice poses ethical concerns. Lawyers must have competent understanding of AI technologies to avoid mistakes that impact their work. Others will say that a refusal to adopt AI technology is a breach of a lawyer’s duty of competent representation. Another consideration is the prohibition against charging unreasonable fees. For example, if a case is document-intensive, manual review may be unethical if an e-discovery option would be more time- and cost-effective. e-discovery also implicates confidentiality interests. Unless a firm has an in-house team that handles document review and electronically stored information, this process requires an outside vendor to store and review material that is oftentimes sensitive and confidential. Attorneys working in these areas must erect proper safeguards to ensure that client confidences are maintained and not disclosed to an adversary or another person/entity.

AI in legal practice is not self-starting or self-running. It requires diligent oversight. Lawyers should not blankly defer to AI technologies. It is their ethical and professional responsibility to avoid this inclination.

Looking ahead

There are distinct benefits to the various AI applications in legal practice. Chiefly, they are vehicles for more efficient practice. That efficiency is highly valued in the evolving work culture, with a growing appreciation for the flexibility of remote work and increased work-life balance. Through increased efficiency, lawyers, especially those of us kneeling at the altar of the billable hour, can focus on billable work instead of wasting time spent poring through stacks of paper that often contain many redundancies. By streamlining billable projects, law offices can better scale and manage more matters simultaneously. There is no question that the most common legal AI technologies (e-discovery platforms, legal research databases) have improved legal practice, are here to stay, and will continue to improve over time. 

AI applications will continue to expand and become more numerous within the legal industry; having said that, it is hard to predict with certainty what a future landscape will look like. 

Fears about job loss are premature. Without plausible AI legal robotics or platforms that can operate without a human trigger, the human element remains critical. Even though certain AI provides a benefit that human intelligence cannot achieve on its own, there is much more to legal practice than data collection and analytics. Outcomes in litigation are often determined by human emotions in settlement negotiations, movement within a courtroom, and facial and bodily expressions to convey a point. For now, the machines cannot compete with humans on those fronts.

Augmented intelligence appears to be the most likely and desired future outcome. This concept blends the best of what humans can offer, with machine intelligence enhancing our limitations. It does not remove the human element. But it acknowledges the immense benefits that AI can bring to legal practice. 

AI and its interaction with the legal industry is a topic of rapid change and expanding intrigue. This article is meant to provide only a brief overview. There are numerous other applications that could have been discussed here. And other applications that are surely in development right now that will require an update to this piece. Be on the lookout.