One of the most pivotal cases of the 21st century is Donald Trump v. Norma Anderson. This case emerged last March when the Supreme Court overturned a Colorado decision, enabling Donald Trump to pursue office once more.
This echoes the Supreme Court's intervention in the Florida recount of December 2000, which ultimately secured the presidency for George W. Bush over Al Gore. Both these cases were delivered behind a veil of anonymity, leaving the public in the dark about the authorship and underlying deliberations.
Recent advancements in artificial intelligence are beginning to pierce this veil, revealing the authors behind some of America's most critical judicial decisions. Approximately 10-15% of the Supreme Court's rulings are issued as unsigned opinions, leaving scholars and observers to speculate about the Court’s inner workings. These opinions, known as per curiam opinions, can carry immense significance.
From defending press freedom to deciding landmark issues like desegregation, unsigned opinions have played a pivotal role in shaping American jurisprudence and history. Traditionally, uncovering the authorship of these rulings has been a slow, labor-intensive process, often relying on posthumous releases of Justices’ papers or educated guesswork based on writing styles.
By leveraging advanced language models and deep learning techniques, similar to those used in identifying Shakespearean works, we can now pinpoint the authors of unsigned Supreme Court opinions with over 90% accuracy.
Who wrote the per curiam opinion in Trump v. Anderson? While speculation points to Chief Justice Roberts or Justice Kavanaugh, in a forthcoming article at the Harvard’s Journal of Legal Analysis our analysis suggests the likely author is Justice Alito, with possible contributions from Justice Kavanaugh. Chances that Chief Justice Roberts authored the opinion are less than 10%. Though this per curiam opinion was presented as representing the entire Court, none of the liberal justices contributed to the writing.
Turning to Bush v. Gore, the 2000 case that decided the presidency, the anonymous author of the unsigned majority opinion has long been a subject of debate. This high-profile 5:4 case propelled George W. Bush to the Oval Office. Scholars have speculated that Justices Kennedy or O'Connor, or possibly both, could be the authors of the majority opinion.
Both justices are conservative voices who refrained from affiliating with any separate opinions. Surprisingly, our algorithm did not identify Justice O’Connor as a potential author of the opinion. Instead, it revealed that Chief Justice Rehnquist likely authored the first part of the majority opinion and that Justice Kennedy is most likely the author of the second part of majority opinion.
The Sebelius case, which upheld the constitutionality of the Affordable Care Act (ACA), saw Chief Justice Roberts—a conservative—join the four liberal justices in a narrow 5-4 decision. But the case also featured an unsigned dissent by the Court's four conservative justices, fueling speculation about its authorship. One particularly intriguing theory claims that Chief Justice Roberts originally wrote the dissent but switched sides, ultimately penning both the dissent and the majority opinion.
Our algorithmic analysis, however, debunks this notion. The data suggest that Justice Scalia likely authored the bulk of the dissent, with Parts IV and V likely penned by Justices Alito and Kennedy, respectively. According to our findings, the likelihood of Chief Justice Roberts having written the dissent is effectively zero. Another conspiracy theory bites the dust.
Some may argue for respecting the justices' desire for anonymity. After all, the tradition of issuing unsigned opinions is intended to emphasize the unity of the Court (or of the majority/dissenting group of justices), and identifying authors could shift focus to individual perspectives, diluting the intended collective message. Notably, in France, a 2019 law even criminalizes data scientists who attempt to predict individual judges’ rulings.
However, the French tradition differs significantly from that of the U.S., where opinions have historically been unsigned and issued in the name of the entire court. In contrast, in the U.S., the norm is to sign opinions.
As AI continues to evolve, reversing these technological advances is neither practical nor desirable. Instead, we should embrace this progress as an opportunity to enhance accountability, transparency, and public understanding of our judicial system and its decision-making processes. Previously unanswerable questions about the Court's dynamics, collaboration across ideological lines, and the evolution of legal doctrines and judicial thought, are all now within a better reach
For at least a century, the Supreme Court has sometimes operated under a veil of anonymity. We are now on the brink of a new era of judicial transparency. It remains to be seen how the Justices will react. Will they conceal their identity? Will they use AI tools like ChatGPT to hinder our ability to identify them?
Ronen Avraham is a law professor at the Univ of Texas in Austin and Tel Aviv Univ in Israel. The piece is based on an academic paper written with Prof. Tamar Kricheli Katz, Prof. Roded Sharan, Rami Nasser and Itamar Kohn.