Can AI pick your judge for the win?

The development and use of Artificial intelligence ("AI") is a daily subject of intense focus these days as new applications for AI regularly appear. This author is published on how AI may someday testify in court, why that testimony would be permissible and not be unduly prejudicial, and why it may become an everyday event once AI starts driving automobiles without human interaction. But why limit AI to the witness stand when we can use it to pick a judge, determine beforehand how a judge will decide a motion, a cause of action or a bench trial, and even predict the outcome of our work? A silly premonition of the future? Or, referring to AI in the plural and as a famous movie once proclaimed, "they're here."

AI has already invaded the judge's headspace and challenges humans to keep up. So far, humans seem to be losing and AI is only improving. Recently, AI created by a team of computer whizzes and legal professionals accurately predicted the outcome of European Court of Human Rights cases 79% of the time. (PeerJ Computer Science "Predicting judicial decisions of the European Court of Human Rights: a Natural language Processing Perspective".) The American Bar Association, as written by the Honorable Herbert B. Dixon Jr. (Ret.), detailed a fascinating study where AI predicted the outcome of Supreme Court cases with greater accuracy than the experts. Humans did well at 59.1%, but AI handily beat them with a 75% predictive accuracy of the court's affirm/reverse results. (ABA, "What Judges and Lawyers Should Understand About Artificial Intelligence Technology".) In criminal matters, AI uses algorithmic predictions of recidivism to better identify bad guys who are more likely to return to a life of crime once released from prison with greater accuracy than human judges. (The Present "Algorithms Identify Repeat Offenders Better Than Judges"). Apparently, the human element of compassion is a hindrance when applied to sentencing guidelines. In short, AI is already predicting what judges will and should do.

What does this mean for trial lawyers who want to win? It likely means that lawyers will soon regularly spend their client's money on AI that will determine where it is best to file a lawsuit, when to file peremptory challenges in state court to replace a judge and how the "wheel" will allocate that judge's replacement. And, specific to the judge assigned, AI will determine how best to proceed in discovery, law and motion, and to and through trial. Predictive AI analysis may even save money. Demurrer attacks will be avoided if AI advises that the assigned judge virtually always overrules them. Motions to Compel Discovery may include sanctions requests, or not, depending on what AI determines the judge likes best. How about a career-long analysis of a judge's preference for certain causes of action. And what about "tweener" scenarios where AI will instruct the trial lawyer whether a judge prefers tort over contract claims, trade secret over patent claims, or common law claims over those that are statutorily-based. Plainly, AI's guidance will surpass old school techniques of viewing comments by grumpy lawyers on various websites grousing about judges who failed to award them a ruling they wanted.

Lawyers are already becoming addicted to AI when representing their clients in court. In fact, some prestigious law firms already use AI lawyers to perform certain work. But reliance on AI may be a hindrance to the trial lawyer when relied upon to prophesize what a judge will do next. The reason is twofold. First, the separation between good trial lawyers and great trial lawyers is often the little things AI simply cannot deliver, like sitting in a judge's court waiting to be called, intently listening to that judge on other matters, and then dialing in the mood on that particular day in order to bring on the best oral argument for a client. Great trial lawyers also build trust with the court. Great trial lawyers work hard on decorum, and they earn the respect of the court. Great lawyering not only defies the dulling predictability that AI promises, but AI cannot understand what great lawyering means and it cannot factor these qualities into an algorithm that makes sense when applied to judges.

Great lawyers also prefer the unknown and dream of what could be, not what is or what was. So, too, great judges do not prefer habit. They enjoy the practice of law as much as great lawyers do and they want to be inspired (when given the opportunity) to make great rulings that AI simply cannot predict. Great judging escapes the predictability that AI promises, and lawyers who settle for AI predictability will never inspire that next judge to render that next great ruling.

It wasn't that long ago in a Los Angeles courtroom that oil companies faced massive consequences after the California Supreme Court decided that zip codes were personal information. The oil companies were immediately sued thereafter for using zip codes at the gas pump in conjunction with credit card transactions. Plaintiff alleged that the oil companies violated a California Act on its face; no escape. But the oil companies used zip codes to protect consumers, not exploit them, and therefore, did not violate the spirit of the Act. AI would have considered the situation, noted the claimed violation of the Act's actual wording, inspected the judge's impeccable reputation for enforcing the law in LA's complex court, and given last rights to the oil companies. But one of the oil companies put its faith in the judge, hoping to inspire him to rule against the tide, to see past the Act's words, and to render a ruling that served justice. It succeeded. AI could never have seen that coming, and if the lawyer representing the oil company had listened to that advice, the motion would never have been brought.

AI is not yet suited (and may never be suited) to analyze how a judge may rule when influenced by brilliant lawyering. AI is helpful but it cannot replace humans in determining trial outcomes; this remains the playground of the human lawyer.

This article is the third in a series of columns focusing on how artificial intelligence is impacting attorneys in and out of the courtroom.