Artificial Lawyering?

-- By AvanequePennant - 20 Feb 2025

The increased prominence and capabilities of artificial intelligence in society have caused anxiety in many sectors and the legal industry has not been an exception. Lawyers have used artificial intelligence to aid in legal research, case strategy, document review, and legal drafting. The most famous examples of lawyering in the age of artificial intelligence illustrate the dangers of these learning models, including the threat of “hallucination,” where lawyers have submitted briefs with citations to make up case law.

While AI has the potential to promote efficiency and access, it also raises concerns about the function and future of human judgment in the legal profession. I understand major concerns about the role of AI in lawyering to fall into two main anxieties: first, that AI will replace the need for human legal expertise, and second, that AI dehumanizes the legal profession.

The Fear That AI Will Replace Human Legal Expertise

AI-powered tools have already reshaped legal research, contract drafting, and document review. Programs like Lexis+ AI, Casetext’s CoCounsel? , and Harvey AI can analyze case law faster than human researchers, draft contracts and other legal documents based on templates, and predict case outcomes using historical legal data. This rapid automation has sparked fears that AI will make some legal jobs obsolete, particularly for entry-level associates and legal researchers.

Despite its capabilities, AI cannot (yet?) replicate the humanity and creativity central to persuasive advocacy and legal judgment. While AI can generate arguments based on precedent, it cannot develop novel legal theories or respond to unexpected courtroom dynamics. Effective lawyering requires emotional intelligence, strategy, and interpersonal persuasion, all of which are essential to achieving favorable outcomes for clients.

The Fear That AI Dehumanizes the Legal Profession

Concerns about AI’s role in law extends beyond fears of job displacement. There is also a broader anxiety that AI will fundamentally alter the nature of legal practice by removing the human element from legal decision-making. While in the U.S. the prospect of substituting the judiciary for technology might seem like a dystopian prospect, countries like China have implemented “internet courts.” In an article entitled “Robot justice: China’s use of Internet courts,” Tara Vasdani described China's 24-hour, 7-day-a-week Internet court system where artificially created judges appear by hologram. These judges set schedules, ask litigants questions, take evidence, and issue dispositive rulings.

As we have discussed in class, the humanity of judicial discretion and reasoning is immensely valuable for case strategy. Firms provide financial incentives in pursuit of clerks who are assumed to have insight into how a judge rules. Especially when unknown factors can carry unknown weights in what a judge does, the possibility of being able to predict is valuable for crafting case strategies. While some may welcome the prospect of bright-line rules and consistent verdicts under an AI judiciary, data-driven models may not (yet?) be able to provide context-specific reasoning and fact-intensive analysis. It is unclear what role, if any, rhetorical devices emphasized in oral advocacy training would be under these data-driven models.

This draft itself could have been written by a bot, couldn't it? It has the same quality of empty abstraction, of names thrown about without actual analysis, the same inflated claims void of philosophical rigor (hedged about with '(yet?)') that characterize the prose that falls to the bottom of the stochastic parrot's cage. Yet, it was written by a human being, already unconsciously tending to imitate the slop being produced by the software that statistically imitates the behavior of the humans now imitating it....

Fake intelligence can't do lawyering. Lawyering is about the conscious understanding and interpretation of social processes based on the consilient overlapping of multiple independent approaches to social understanding. It requires exactly those qulities that conclusively distinguish human intelligence (which creates infinite understandings of what must and cannot be true about the world from tiny amounts of imperfect information) from statistical pattern-matching that uses vast amounts of information to achieve fragile guesses about what might be true, without any method for conscious reflection on the outcomes. The point of this course was to exemplify that distinction. The draft would become stronger by (1) actually explaining what "AI" is, and (2) putting itself in contact with the argument I spent 14 weeks building that explains why it is of no significance at all to the activity of lawyering.

---

You are entitled to restrict access to your paper if you want to. But we all derive immense benefit from reading one another's work, and I hope you won't feel the need unless the subject matter is personal and its disclosure would be harmful or undesirable. To restrict access to your paper simply delete the "#" character on the next two lines:

• #Set ALLOWTOPICVIEW = TWikiAdminGroup, AvanequePennant
• #Set DENYTOPICVIEW = TWikiGuest

Note: TWiki has strict formatting rules for preference declarations. Make sure you preserve the three spaces, asterisk, and extra space at the beginning of these lines. If you wish to give access to any other users simply add them to the comma separated ALLOWTOPICVIEW list.