AI: The New Kid in Jury Consulting
- Elizabeth Babbitt, M.A.
- May 26
- 5 min read
Updated: Jun 11

Elizabeth Babbitt, M.A.
Jeff Dougherty, M.S.
My reason for being in trial consulting comes down to the saying, “trials are about important events in the lives of people.” Put simply, this business is a business of people, not machines. But if you’ve been paying attention lately, you’ve probably heard the hype about artificial intelligence (AI) and, more specifically, large language models (LLMs). Are these just the latest buzzwords, or do they actually have a place in the gritty, real-world work of jury research?
Let’s break it down—no jargon, just the facts.
What’s an LLM, Anyway?
Picture this: an LLM is like the world’s most well-read legal expert. It’s been trained on mountains of text - books, articles, case law, you name it - and it has learned through all this data to predict the “next best word.” So, LLMs aren’t actually “artificial intelligence” in the truest sense; they don’t “think” or “reason” like a person. But they are great at spotting patterns and generating text that sounds pretty convincing.
Today’s LLMs, like OpenAI’s GPT 4.1 and o3, Google’s Gemini 2.5, Anthropic’s Claude Opus 4, and others, can:
Spit out 100 draft voir dire questions faster than you can refill your coffee
Summarize a stack of case documents in seconds and even produce a podcast that you can query about what it’s read
Suggest new ways to frame your case theme and provide specific feedback
Role-play as a mock juror (with some big caveats)
But before you start thinking you’ve unlocked the secret to jury research and trial science, let’s talk about what these tools can, and can’t, do today.
Where AI Can Actually Help
Here’s where AI demonstrates its strengths for trial attorneys and consultants:
Brainstorming: Need fresh angles for your opening? Want to wordsmith theme ideas? Stuck on the best voir dire questions to weed out juror bias? AI can throw out a dozen ideas in less time than it takes to order lunch.
Drafting: First drafts of juror questionnaires, witness outlines, or closing arguments? AI can get you started, so you can focus on the real strategy.
Efficient Document Review: AI can quickly sift through mountains of documents, flagging inconsistencies, contradictions, or key facts to address with witnesses.
Challenging Blind Spots: Sometimes, we all get stuck in our own heads. Moving with inertia along the path we’ve always traveled. AI can help you see arguments or perspectives you might have missed.
Think of AI as your brainstorming buddy, the best mentee intern you ever employed; never tired, never out of ideas, but still needing your input and judgment to separate the gold from the fool’s gold.
Where AI Falls Flat
It cannot be stressed enough - AI is not a substitute for real people, real experience, real human emotion or attitudes, or real-world testing. Here’s why:
Risk of “Hallucinations”: AI sometimes fabricates information or citations, often with such confidence that it can be unnerving to question - even for the best experts. Any AI-generated content must be carefully checked for accuracy to avoid undermining credibility or violating ethical rules. For example, as of this writing, there have already been situations in which attorneys have been sanctioned for citing case law for cases that AI fabricated out of thin air.
No Human Intuition: AI doesn’t get nervous or experience emotions, so even though its words may make people read and interpret feelings, it cannot empathize with its human “overlords” as it does not have feelings. It does not pick up on subtle body language or emotional undercurrents that can make or break a case. This is a problem on two fronts when it comes to witness testimony. First witness preparation: AI cannot identify, ,and resolve emotional barriers to effective witness testimony like a human can, which is a vital component of effective witness preparation. Second, AI has never interviewed actual jurors and doesn’t know how jurors assess witnesses within the context of adversarial and emotional questioning. Therefore, it cannot teach witnesses how to testify in a manner that will garner juror trust in this unique communication environment.
Biases in, Biases out: If the data the model learned from is biased (and it always is), AI can spit out some pretty tone-deaf, problematic, or even unethical suggestions.
Ethical and Legal Risks: Confidentiality, data privacy, and transparency about AI’s role in preparation are critical. Over-reliance on AI can open the door to mistakes and ethical pitfalls.
No Field Testing: The best jury research is still boots-on-the-ground; focus groups, mock trials, post-trial interviews with real human beings. AI can’t replicate that. It is an undisputed fundamental tenet of research methodology that participants must “mirror” as closely as possible the population one is studying. This is why we do not test jury populations in Texas for cases that will be tried in front of juries in the Southern District of New York.
Bottom line: AI can help you brainstorm, but it can’t tell you how a real juror from your specific jurisdiction is going to react to your case and it can’t replace the value of human expertise when judgement calls are needed using the entire human perspective in the room.
So, Should You Use AI? (Short Answer: Yes, But…)
In the words of AI researcher Ethan Mollick back in June 2023, “Today’s AI is the worst AI you will ever use.” AI is here to stay, and it’s only going to get better. It’s time to learn to become a “human in the loop.” Keep in mind, however, using AI is like having a sophisticated lock-picking tool for a safe: it can help you open the safe faster and more efficiently, but it still takes a skilled locksmith to know which safe to open, when to stop, and how to handle what’s inside responsibly. Use it to:
Jumpstart your thinking when you’re stuck. AI is great at refining a large amount of information to its essence, even in a thematic fashion. This can be very helpful to “get the juices flowing” when brainstorming how to organize an opening statement for example.
Draft the basics so you can spend more time on the tough stuff. For example, you might have numerous themes you want to communicate, but you’re struggling to organize them in a coherent and cogent manner. AI can do the heavy lifting and brainstorm numerous approaches you might not have ever considered.
Help organize and review alongside your own deep review so you can check your memory quickly and push through quick summations so you can start ahead of the game and spend less time in the administrative weeds.
Pressure-test your arguments and spot blind spots, even logical inconsistencies in your, and your adversary’s arguments you had either missed, or never considered. Tell your AI to assume the persona of your opponent, ask AI to argue with you so you can start looking around corners and be sure your view is not becoming overly narrow.
But don’t let AI be the final word.
The Verdict
AI is a powerful new tool in the trial attorney’s and consultant’s toolboxes. But just like you wouldn’t send a rookie attorney to deliver your closing, you shouldn’t rely on AI to be the ace on the team. Use it to get smarter, faster, and more creative, but always double-check its work with your own expertise, seasoned behavioral scientists, ethics, code of conduct, and real human juror feedback. There is no machine replacement for tasks where judgment and experience are needed.
The future is bright and certainly going to look very different than the days of post-it notes everywhere and endless stacks of files lining our office walls. With the right mix of technology and human judgment, we can keep helping our clients through these important events and see their cases through the eyes of the people who matter most: the jury.
By Elizabeth Babbitt, M.A.
Principal Jury Research Consultant
Insights Alignment, LLC
and
Jeff Dougherty, M.S.
President and Senior Jury Consultant
Litigation IQ
Comments