Business of Law

AI tools can help litigators pick jurors who might be on their side

  • Print

Image from Shutterstock.

A jury is considered the gold standard of U.S. law and a symbol of justice designed to protect the innocent and prosecute the guilty in a fair, impartial trial. And now, artificial intelligence tools are available to help attorneys identify prospective jurors favorable to their cases. But do jury trials remain golden if one side uses AI to select the perfect jury to win?

Companies—including Magna Legal Services, Momus Analytics and Vijilent—use AI to scour the internet and capture and filter vast amounts of information about potential jurors. It would be nearly impossible for lawyers to cost-efficiently find and review that data on their own within the time constraints of voir dire.

AI can also help jury consultants and lawyers analyze case information, biographical data from jurors’ questionnaires and community surveys to help pick jurors. Providers use web-scraping bots to collect text from social media and websites and then use machine learning and natural language processing tools to analyze all collected data and extract information relevant to prospective jurors and useful in voir dire. Data analytics and visualization tools surface the data for litigants to review, allowing them to frame the right questions and get the correct answers to challenge jurors for cause or use peremptory strikes.

Magna’s JuryScout offers social media surveillance searches on prospective jurors. Vijilent’s Reveal service obtains comprehensive reports highlighting potential jurors’ hidden biases and risks through social media and public records analyses.

Momus Analytics claims its technology can help attorneys identify and rank a case’s best and worst jurors. Its software uses a proprietary scoring algorithm and analytics engine to score potential jurors. It says it does not use illegitimate factors like age, country of origin, race, religion or sex to determine ratings. Momus’ software also compiles and analyzes jurors’ preferences and combines them to determine which particular jurors will give an attorney the best chance of winning. The software predicts how individual jurors may impact the group’s decision-making process.

“AI is wonderful for data gathering,” says Richard Gabriel, the president of Decision Analysis, a trial consulting company based in Los Angeles. “Where I start having real questions has to do with interpreting or producing an evaluative rating. You can develop some useful information from your data gathering.” For example, “There are components of leadership you can get from the status of someone’s employment, whether they are a manager or university professor, but it does not tell you about their personality, which is the main component of leadership that you’re looking for.” Gabriel says. “Sometimes in jury selection, you’re looking for the person who talks the most in voir dire or is the most sociable, not necessarily the person with the greatest professional accomplishments.”

“Data mining large data sets is the new wave of AI analysis,” says Daniel Wolfe, senior director of jury consulting at Magna. “Advanced tools provide more sophisticated data mining more quickly, and that’s the challenge with social media and background searches—there’s such a wealth of data out there, and in most cases, we don’t get the list of jurors until the day of” the trial.

Still, prospective juror information is electronically fed into the “platform that helps analyze, organize and report the data quickly for the client,” Wolfe says.

Survey says

A jury selection survey typically focuses on the case at hand and includes questions to assess a synopsis of the case and a summary of the evidence; jurors’ attitudes toward the specific case and the legal system; their demographic data; and their thoughts on defendant culpability. The survey remains a primary tool of scientific jury selection to collect and weigh community sentiment on trials and understand potential jurors’ biases, likes, dislikes, habits and social backgrounds.

Surveys adopt modern tools to portray cases to respondents. Wolfe says Magna surveys a large segment of the population from a trial venue to engage in quantitative analysis. Potential jurors watch a self-paced summary presentation of the evidence and are asked to provide feedback on a likely verdict and damages. An algorithm runs simulations and provides consultants and clients “robust statistical data on damages,” Wolfe says.

Ashwin Murthy, an artificial intelligence engineer and founder of the California-based startup Jurypicks, says AI tools can help attorneys train in jury selection by using simulated data sets gathered from surveys.

Jurypicks feeds survey information into a data model looking for statistical patterns. For example, in a predefined wrongful termination case, Jurypicks would look at the propensity of a juror—such as a person who has a higher education degree or owns a business—to support a plaintiff or defendant. Then it would use two or more attributes or features to analyze the percentage of jurors supporting either side. Its Voir Dire Simulator is designed to improve jury selection by allowing attorneys to practice voir dire on a library of cases. At the same time, the Jury Analytics product uses AI to quantitatively analyze and identify the most compelling questions for prospective jurors. The Focus Group allows customers to conduct personalized studies to determine a specific case’s most effective voir dire questions and jurors’ perceptions of them. Jurypicks expects to have a minimum viable product later this year.

In-person focus groups and mock trials use real-time feedback mechanisms to learn how jurors respond to case presentations.

“AI adds another layer of analytics to the coding and interpretation of how jurors respond,” Wolfe says. “The AI tools help identify the significant peaks and valleys when jurors shift in their perceptions on rating the presentations.”

Unfair advantage?

Many states have no rules governing the procedure for voir dire and the permissible scope of the process. Generally, the manner and content of voir dire are left to the trial court’s discretion and the bounds of relevance to the parties or issues in the particular case. For example, a party could find out about a prospective juror’s favorite brand or flavor of ice cream if that were relevant to the case.

So can social media research predict juror outcomes? “Absolutely not. Lifestyle characteristics may give you an indication of which way a juror may lean or interpret some issues in the case but cannot give you the whole person. We’re always looking for life experience, attitudes and beliefs. Some of that you might be able to glean from social media posts—whether tweets, Facebook posts or TikTok,” Gabriel says.

“The question is, how true is that to who the person is?” he adds. Because while “people post a great deal about themselves online about how they would like to appear to the world, it doesn’t necessarily have to do with their decision-making criteria, psychological makeup and how they’re going to interact with a group.”

Although AI appears to be the next step in scientific methods of jury selection, Wolfe says, “Most of the time, the client wants the data interpreted and strategic counsel. AI and advanced tools are designed to interpret much of the data, but there’s still the layer of insight and intuition that only comes through experience.”

If litigants use AI tools to select jurors, they must still drive their selection strategies within the confines of judicial discretion and the procedural limits of voir dire. And if an attorney used AI to choose a winning jury, was it a strategic advantage, and did AI win the case for them? The jury is out on that.

This story was originally published in the April-May 2023 issue of the ABA Journal under the headline: “Stacking the Deck: AI tools can help litigators identify and pick jurors who might be on their side.”

Give us feedback, share a story tip or update, or report an error.