Wile E. Coyote and the rocky, stupefying, future-is-now of artificial intelligence-enhanced litigation

By Jack Ucciferri
Smiley faceThe Colleges of Law - Fri, Feb 09, 2024

In this edition of Legal Currents and Futures, The Colleges of Law continues sharing a series of thought pieces about artificial intelligence and the law.

Society is amid a collective Wile E. Coyote moment. 

Society is now suspended in thin air—having raced off an entirely foreseeable technological cliff, legs spinning as if solid ground was still beneath our feet. Things are about to change fast. There is good reason to hope that recent advances in machine learning and artificial intelligence (AI) may result in unimaginable benefits for humanity. Alternatively, we could experience a collective face-first free-fall into a deep canyon, or likely some combination of both. In any case, it seems clear that adaptation to new realities wrought by artificial intelligence will not be optional for responsible professionals.  

THE FUTURE IS NOW

Attorneys staying abreast of advances in machine learning and AI seem to share a “Wow, the future is now!” feeling. Here are some future-is-now examples of how some tech-savvy litigators already incorporate machine learning and AI into their practice. 

1. Document Review and Analysis

ML/AI’s extensive analysis of documents has revolutionized modern document review. ML/AI-powered tools can sift through millions of pages of text, extracting nuanced information that can inform decision-making. Moreover, these tools operate via natural language in a conversational style, including concepts and even jargon. Lawyers may tell a well-trained AI bot to “find me scorching hot docs in the attached batch of emails regarding the pertinent issues we have previously discussed. Please look for instances involving attempts to communicate these issues using code words or indirect language.” Boolean search will soon be passé.

The acceleration of legal research capabilities through ML/AI-driven tools has been impressive over the past several quarters. ML/AI can scour massive databases for case law and regulations in a fraction of the time it would otherwise take using tools that many would still consider relatively advanced. The most cutting-edge tools will synthesize legal research into a narrative or recommendation. Will this result in less demand for junior associates, junior associates spending more time doing higher-level work, or both? We are about to find out. 

3. Predictive Analytics

ML/AI’s predictive algorithms can transform risk assessment and case strategies. Historical case data can now provide insights to aid litigators in determining the value of contingency-based cases, managing client expectations, and guiding strategic decisions.

4. Litigation

AI’s ability to categorize documents based on responsiveness has streamlined the discovery process. Generative AI can draw upon an entire case file to suggest pertinent deposition questions, interrogatories, or requests for admissions. Generative AI can also suggest discovery responses relevant to the facts of the case and prior meet and confer efforts. Of all the areas of litigation, e-discovery companies have moved quickly to adopt generative AI within their current offerings.

5. Drafting

Imagine a ChatGPT-like AI application trained exclusively on millions of pages of legal writing that does not hallucinate and includes mechanisms (links, presumptively) to allow for verification of assertions. Litigators are beta-testing that technology through Westlaw, Lexis, and smaller boutique legal technology start-ups focused on specific legal areas like personal injury. We will all be able to buy such tech within a year. Are firms ready for that? No.

THE FUTURE IS STUPEFYING

“Stupefied” reflects a sentiment felt by some who spend time thinking about how technology shapes the future of law. The next section lists examples of how ML/AI may change legal practice. Some of the following examples are predictions, some mere speculation. As this goes to press, many other exciting use cases for ML/AI are being imagined and developed.  

Virtual assistants play an important role in legal operations. AI-powered virtual assistants will improve dramatically in the next few years. Administrative tasks such as scheduling, intake, document assembly, filing, and even marketing will become increasingly handled via machines. Human attorneys will interact with these AI assistants in ways resembling how they interact with their assistants today. These digital assistants will be able to handle their assignments so efficiently that they will sit idle most of the time. The most profitable firms will have thorough processes and prompts for getting the most out of their digital assistants.

2. Judicial Customizations

Today’s judges must read interminable pleadings of highly variable writing quality, formatting, and tone. On the other hand, litigators often need to be more knowledgeable about the preferences of any given judge regarding writing styles in their pleadings. Following the trend of e-discovery, each judge might suggest or even mandate that all pleadings flow through a large language model that offers edits according to the judge’s individual preferences. Alternatively, private sector companies will sell products purporting to provide this service. 

3. Artificially Enhanced Judicial Logic

This idea is far out there but stay open-minded. Assuming AI keeps advancing at the rate it has been over the past several years, from an objective standpoint, it would be arguably unethical for judges and juries not to incorporate AI analytics into the adjudication process. Algorithms drawing upon billions of data points could help fact finders better differentiate compelling legal arguments and evidence from merely charismatic courtroom performances and precedential cases. AI cannot, should not, and will not replace human fact finders.  

4. Emotion and Sentiment Analysis

Did the judge sigh, or are they holding back a sneeze? Is juror number seven frowning, or is that what their face looks like when listening intently? Are the deponent’s eyes dilated because they are nervous or high on a narcotic? Soon—maybe  even now—AI can analyze such proceedings for emotional cues that could empower lawyers to adapt tactics during a trial. Of course, there are currently rules restricting the use of such technologies, and there would also be ethical issues with the use of such technologies. Like it or not, though, governing bodies will likely be forced to confront the use of such technologies in the coming years. 

5. Real-Time Body and Language Analysis

The judge just made remarks while ruling on a motion in limine. Alternatively, opposing counsel just made a speaking objection. However, what does any of that mean? Did the judge’s extemporaneous remarks reveal a bias toward hearing more from a particular witness or against a particular line of reasoning? Was the opposing counsel’s speaking objection poorly worded or an attempt to distract the jury? We mere humans could only guess—and if we are being honest, we cannot generally muster the brainpower to even ask ourselves those sorts of questions in the heat of the moment. However, if allowed to be used in these settings, the AI co-counsel of tomorrow could both ask and answer those questions and then provide guidance on how to adjust tack accordingly. 

Of course, as we consider AI’s promise, we must also confront its challenges.

1. Data Bias

The significance of data bias in the training of AI models is a significant and poorly understood risk. As AI models mesh into programs that inform legal proceedings, there must be mechanisms to ensure the fairness and impartiality of the datasets and the trainers of those AI models. Given the fundamental biases reflected in the corpus of centuries of writing and human nature, this challenge will be difficult to overcome. 

2. Balancing AI and Human Judgment

As institutions of every type accelerate the incorporation of AI into their workflows, leaders will struggle to balance the perspectives of human judgment with AI’s efficient decision-making. High-profile mistakes will be made, even by the most competent leaders. 

3. Data Security

Data security is already complicated to manage. AI is in the process of escalating that challenge exponentially. AI programs in circulation are already capable of cracking passwords and finding/exploiting network vulnerabilities. Of course, other programs provide previously unimaginable power to those seeking to keep our data safe. I am glad data security isn’t my job. 

Law schools and their accreditors will fail to adapt mainstream pedagogy on the required time scale (immediately). Recognizing that their law schools cannot keep up with technological advances, digital-native law students will teach themselves the AI-related skills they will need to thrive in their future workplaces. Integrating the next generation of young attorneys into the legacy workflows of established firms will be challenging for all parties. 

TECHNO TAKEOFF OR WILE E. COYOTE-STYLE FREEFALL: THE CHOICE IS YOURS 

Whether one accepts that we are running on thin air, AI is poised to upend law practice relatively quickly. Those who take the tremendous power of ML/AI seriously will be positioned to thrive in the future—those who do not risk hitting the bottom of a canyon in a cloud of dust.  

Jack Ucciferri is an associate with Snyder Burnett Egerer LLP. He is pursuing a self-guided legal tech and AI education in his spare time. Sometimes, he even has dreams about it; sometimes, those dreams are nightmares. Among other things, Jack has worked on training large-language models for legal applications, engineering high-value prompts for generative AI bots to complete legal tasks, and saying yes to every offer for a demo of a legal tech product he receives. You can reach him at [email protected].