AI Brings Supreme Court Decisions to Life

The U.S. Supreme Court has long been an institution defined by tradition and deliberate resistance to change. Yet an innovative project is now using artificial intelligence to break down one of the most significant barriers to public access: the inability to see and hear justices announce their decisions from the bench.[1]

For decades, the Supreme Court has maintained a curious restriction on its most dramatic moments. While oral arguments have become increasingly accessible—especially since the COVID-19 pandemic made them routinely broadcast—the justices’ bench announcements of decisions remain locked away until months after they are rendered.[1] Only those physically present in the courtroom can witness the moment a justice summarizes a ruling or delivers an oral dissent. This creates an unusual gap in public access to one of the government’s most consequential institutions.

The Vision Behind On The Docket

Northwestern University professor Jerry Goldman has been working to democratize Supreme Court access since 1996, when his nonprofit project Oyez launched online to provide audio recordings of oral arguments and opinion announcements dating back to 1955.[1] Now, Goldman’s team is taking the next step with a new independent project called On The Docket, which uses AI to generate visual depictions of the justices speaking words they actually spoke in court.[1]

The motivation is straightforward. As Goldman explains, “Since it’s public in the courtroom, it should be public for everybody. That’s simple.”[1] Yet implementing this vision proved technically complex. Without cameras allowed in the Supreme Court, the team needed an innovative solution to create the visuals that courtroom observers experience.

Creating Realistic Avatars

The answer: AI-generated avatars of the justices.[1] Working with University of Minnesota professor Timothy R. Johnson and Spooler, an AI design company, the project team faced significant technical challenges. Johnson recalled that early AI attempts produced “hilarious” results, with robots magically making justices disappear from the bench or having all of them bend forward simultaneously.[1]

The breakthrough came through leveraging existing public materials. By using photographs and videos from the justices’ public appearances, the team created realistic video versions of each justice, capturing their distinctive mannerisms, head tilts, and hand gestures.[1] These avatars were then synchronized with the actual audio from the courtroom, creating a complete visual record of decisions that previously only existed as text or audio.

Navigating Ethical Questions

The project’s developers confronted important ethical considerations. Should they make the video indistinguishable from reality, or should they signal to viewers that the visuals were artificially generated? The team chose a middle path: slightly cartoonizing the video while clearly marking it as AI-generated.[1] This approach ensures viewers understand what is real—the audio—and what is not—the video.

This transparency is particularly important given the Supreme Court’s historical resistance to cameras. By being explicit about the AI-generated nature of the visuals, On The Docket respects the Court’s concerns while still expanding public access to these crucial moments.

The First Test Case

The project’s inaugural effort demonstrates its potential. The team created a visual of Chief Justice John Roberts’ 14-minute summary from the bench of the Court’s 6-to-3 decision granting former President Trump and all future former presidents complete immunity from prosecution for core official actions taken while in office.[1] This landmark ruling, previously available only to those in the courtroom, can now be experienced by the public through the AI-generated visuals paired with the actual audio.

A Decades-Long Campaign for Access

The Supreme Court’s restrictions on bench announcements have frustrated journalists and scholars for years. Reporters and legal experts have repeatedly requested that opinion announcements be made available for live broadcast, but these requests have been consistently denied.[1] Interestingly, Goldman notes that he has reviewed papers from the early Warren Court in the 1950s in which justices discussed memorializing oral arguments and opinion announcements through tape recordings, with “no indication in those papers that they wanted to keep these secret.”[1]

The COVID-19 pandemic did lead to one breakthrough: oral arguments are now routinely broadcast, making the Court’s proceedings more transparent than ever before.[1] Yet the bench announcements—arguably the most dramatic and important moments in the Court’s work—remain unavailable until the following term.

The Broader Implications

On The Docket represents a creative solution to an institutional constraint. Rather than waiting for the Supreme Court to change its policies, the project uses cutting-edge technology to provide the public with access to information that is technically public but practically inaccessible. The AI-generated avatars preserve the integrity of the actual audio while making the complete experience available to anyone with internet access.

As artificial intelligence continues to transform various sectors, its application to Supreme Court decisions exemplifies how technology can enhance democratic access to government institutions. By making the drama and gravity of Supreme Court decisions visible to the broader public, On The Docket fulfills a fundamental principle: that the workings of the highest court should be transparent to the people it serves.

Original source: NPR News – AI brings Supreme Court decisions to life