The University of North Carolina School of Law tested whether artificial intelligence (AI) could think like U.S. Supreme Court justices. The results were notable. The AI justices reached the same decision as the real court in a test case.
However, the experiment also revealed limits. The AI systems could follow legal arguments, but they struggled with natural courtroom exchanges.
The test shows how AI could help in legal education. It may also help lawyers prepare for court and practice answering judges’ questions.
UNC Law Tests AI Supreme Court
On October 22, 2026, students and faculty at the University of North Carolina School of Law held a mock Supreme Court hearing. Three AI justices took part in the exercise.
The event, called
May It Please the Machine, tested how well AI could follow legal rules and reach a decision.
The team created three AI justices. Each used a different approach to the law: textualist, pragmatic, or progressive.
The systems used ChatGPT to take part in the mock hearing. Organizers wanted to see how well AI could apply legal rules, weigh arguments, and decide a case.
Assistant Dean for AI Nicole Downing said the goal was not to replace judges. Instead, the team wanted to learn whether AI could follow the reasoning found in real court opinions.
AI Matches Supreme Court Ruling
The Case Behind the Experiment
The mock hearing focused on
Monsalvo Velázquez v. Bondi. The Supreme Court decided the immigration case in 2025.
The case concerned a 60-day deadline for people ordered to leave the United States voluntarily. The main question was simple: What happens when the deadline falls on a Saturday?
The dispute asked whether the deadline should move to the next business day. Organizers chose the case because it raised a clear legal question.
To prepare the AI justices, university data librarian Rolando Rodriguez provided key legal materials. These included court filings, legal briefs, statutes, court rules, and agency guidance.
He also supplied important court opinions. In addition, each AI justice received opinions from the real justice whose legal approach it was meant to reflect.
This gave each AI system a distinct view of the case.
AI Reaches the Same Result
The real Supreme Court ruled 5-4 in favor of extending the deadline to the next business day. This rule applies when the deadline falls on a weekend or holiday.
Justice Neil Gorsuch wrote the majority opinion. Justices Clarence Thomas and Samuel Alito dissented.
The AI justices reached the same result during the mock hearing. The pragmatic and progressive systems formed the majority. Meanwhile, the textualist system dissented.
This result suggests that AI can match a court’s decision when given the right legal materials and clear instructions.
However, one case cannot prove that AI can solve every legal dispute. More tests would be needed to assess how well these systems handle other cases.
AI Struggles in Courtroom Exchanges
Although the AI justices reached the same result, their courtroom performance showed clear limits.
The systems asked questions and raised possible scenarios for the law students. However, their exchanges felt less natural than a real Supreme Court hearing.
Human justices often interrupt lawyers and challenge their claims. They may also change direction when an answer raises a new issue. The AI justices struggled to match this style.
Downing said this was an important lesson from the test. AI can copy parts of legal reasoning, but it does not yet act like a human judge in a live hearing.
This highlights a key difference. Reaching a sound legal result is not the same as taking part in a complex courtroom debate.
For lawyers, that difference matters. They must do more than cite cases and explain the law. They also need to answer tough questions, adjust their arguments, and respond to judges in real time.
How AI Helps Lawyers and Students
AI tools are becoming more common in legal work. Many tools can review court opinions, analyze legal issues, and help users prepare arguments.
Meanwhile, law students can use AI to practice moot court hearings and oral arguments.
The UNC Law experiment points to several ways these tools could support legal work.
Preparing for Oral Arguments
AI could help lawyers find weak points in their arguments before a court hearing.
For example, a lawyer could ask an AI tool to act like a judge. The tool could then challenge the lawyer’s reading of a law or question a key claim.
Lawyers could use these questions to improve their answers. They could also prepare for difficult exchanges before entering the courtroom.
Similarly, law students could practice answering questions from judges with different views of the law.
Understanding Judicial Decisions
AI can also help lawyers study how judges read laws and apply past rulings.
By reviewing court opinions, AI tools can raise questions and compare different legal views. They may also help lawyers spot issues that could affect a case.
However, lawyers must check all AI-generated legal claims against trusted sources. They should never assume that an AI tool can predict how a judge will rule.
Improving Legal Education
Law schools can use AI tests to teach students how to review machine-generated legal work.
For example, students can compare an AI answer with the court’s actual opinion. This helps them spot errors, missing facts, and weak arguments.
As a result, students can build stronger research and reasoning skills. They can also learn when AI is useful and when it falls short.
AI Needs Human Legal Judgment
The experiment does not show that AI can replace judges. Instead, it shows how AI may help with specific legal tasks.
Judges do more than match facts to legal rules. They must weigh arguments, read laws, and apply past decisions to new disputes.
Lawyers also rely on experience and judgment. They must understand their clients’ needs and choose the best way to present a case.
AI can support this work. However, lawyers must check its answers, confirm legal sources, and judge whether its conclusions make sense.
The test also shows why accuracy alone is not enough. An AI system may reach the right result but still struggle to explain or defend that result in a live hearing.
AI’s Impact on the Legal Industry
The UNC Law experiment offers a useful look at AI’s role in the legal field.
The results show that AI can follow some forms of legal reasoning when given relevant materials and clear instructions. However, it still struggles to match the quick exchanges of a real courtroom.
For law firms, this matters when choosing tools for legal research, case planning, and oral argument practice.
Law students may also gain from using AI to test their arguments. However, they must continue to build their own legal skills. Employers may value candidates who know how to use AI while checking its work.
Overall, AI could become a useful tool for lawyers and law students. It can help them prepare, test ideas, and review legal arguments. However, human judgment and strong advocacy remain vital to the justice system.
Explore legal career opportunities and find your next job with
LawCrossing.
See Also:
Law Firm Cyberattacks Surge as Hackers Exploit AIThe post
AI Supreme Court Tested for Legal Accuracy by Law School first appeared on
JDJournal Blog.
