Researcher

Jed Stiglitz

Also known as Edward H. Stiglitz

Jed Stiglitz is a Cornell law professor whose research connects language models with legal reasoning. He earned degrees at Pomona and Stanford, clerked for a judge on the District of Columbia Circuit, and coauthored a book on elections and wrote a book on the administrative state. His 2023 paper focused on language-model annotation of legal reasoning, and his 2025 work mapped data curation, annotation, and verification for legal AI. In 2026, he introduced a transformer model for predicting Supreme Court decisions. He directs Cornell's Center for Law and AI and serves as Associate Dean for Academic Affairs and Richard and Lois Cole Professor of Law.

Works
68
Citations
201
h-index
7

Publication numbers from OpenAlex, as of 27 October 2026.

Timeline

  1. 2002–2012

    Pomona and Stanford degrees

    He earns a B.A. at Pomona College in 2002, an M.A. at Stanford University in 2009, a Ph.D. at Stanford University in 2011, and a J.D. at Stanford Law School in 2012.

  2. 2012

    The Reputational Premium

    He coauthors The Reputational Premium with Paul Sniderman, a book on party identification and policy reasoning published by Princeton University Press in 2012.

  3. 2022

    The Reasoning State

    Cambridge University Press publishes The Reasoning State in 2022. The book links trust, reasoning, and procedural regularity to the growth of the administrative state.

  4. 2023

    Modeling legal reasoning

    He coauthors the EMNLP 2023 paper Modeling Legal Reasoning: LM Annotation at the Edge of Human Agreement.

  5. 2025

    Tasks in legal AI

    In 2025, with Allison Koenecke, David Mimno, and Matthew Wilkens, he identifies data curation, data annotation, and output verification as three areas of relevance to legal AI practitioners. The paper describes ways to curate legal texts, support expert annotation, and evaluate outputs.

  6. 2026

    The Predictable Court

    In 2026, he introduces a fine-tuned transformer trained on fifteen years of unstructured Supreme Court data to predict judicial outcomes. Tested against OT2024, the model reaches 74 percent vote-level and 82 percent case-level accuracy.

  7. By 2026

    A D.C. Circuit clerkship

    By 2026, his career includes a clerkship with Judge Stephen F. Williams of the D.C. Circuit after his Stanford Law School J.D. in 2012.

  8. As of October 2026

    Law and AI center director

    As of October 2026, he is Cornell's Associate Dean for Academic Affairs and Richard and Lois Cole Professor of Law, and directs the Center for Law and AI. The Center's leadership connects law, technology, ethics, and public policy with machine-learning expertise from across Cornell.

Career

  1. NowCornell UniversityRichard and Lois Cole Professor of Law
  2. NowCornell UniversityAssociate Dean for Academic Affairs
  3. NowCornell UniversityDirector, Center for Law and AI
  4. United States Court of Appeals for the District of Columbia CircuitLaw Clerk

Education

  1. 2012Stanford UniversityJ.D.
  2. 2011Stanford UniversityPh.D.
  3. 2009Stanford UniversityM.A.
  4. 2002Pomona CollegeB.A.

Sources and credits

Photo: We have no published photo of Jed Stiglitz yet. Send us a free-licensed photo.

  1. Permanent Archives - Page 3 of 7 - Cornell Law School lawschool.cornell.edu. Checked 27 October 2026.
  2. Jed Stiglitz - Cornell Law School lawschool.cornell.edu. Checked 27 October 2026.
  3. The Reasoning State cambridge.org. Checked 27 October 2026.
  4. Jed Stiglitz jlexcode.github.io. Checked 27 October 2026.
  5. [2504.01349] Tasks and Roles in Legal AI: Data Curation, Annotation, and Verification arxiv.org. Checked 27 October 2026.
  6. The Predictable Court – SIOE 2026 sioe.org. Checked 27 October 2026.
  7. Light Speed: Cornell Law Opens a Cross-Disciplinary Center for Law and AI – FORUM forum.lawschool.cornell.edu. Checked 27 October 2026.
  8. Jed Stiglitz | Department of Government government.cornell.edu. Checked 27 October 2026.
  9. About Us – Center for Law and AI sites.lawschool.cornell.edu. Checked 27 October 2026.