U.S. Supreme Court - The FindLaw U.S. Supreme Court Opinion Summaries Blog

AI Can Predict Supreme Court Decisions, New Study Finds

When it comes to predicting decisions by the U.S. Supreme Court, talking heads may soon be replaced by talking bots on the evening news.

That's because studies reveal that artificial intelligence predicts the court's decisions more accurately than legal experts. In the 2002 term, an algorithm correctly predicted the court's decisions 75 percent of the time. Rolling back 100 years, a machine-learning model correctly predicted 70 percent of some 28,000 decisions.

With the usual legal pundits coming in at 66 percent, it may be time for legal analyst Dan Abrams to make room for Siri, Esq.

Inside the Court

"Every time we've kept score, it hasn't been a terribly pretty picture for humans," says Daniel Katz, lead author of the latest study and a law professor at Illinois Institute of Technology in Chicago.

The new study used the Supreme Court Database, which contains cases dating back to 1791. The researchers drew on 16 features of each vote, including the justice, the term, the issue and the court of origin.

With the information, the research team created a machine-learning statistical model to look at all prior years for associations and outcomes. The algorithm then took the information to update its strategy and progress each year with outcomes.

Scientists in past studies said the new algorithm is "rigorous and well done," while "curating really large data sets and using state-of-the-art methods."

Outside the Court

Katz said the program could be valuable to law firms and their clients because of the cost of pursuing a case to the Supreme Court. With a tool to predict results, they may find it more practical to pursue other strategies. It could also help them tailor their presentations to the justices.

Jazz Shaw, writing for Hot Air, joked that smart devices could even replace the judges. While some lawyers are concerned about losing their jobs to smart robots, perhaps the Supreme Court should think about the future, too.

"Cut out the middle man (or middle woman as the case may be) and just automate the process," Shaw said. "We're supposed to be cutting back on staffing to save money anyway and the robots will work for free while processing a lot more cases."

Related Resources: