Crosby Intelligence Fellowship

Service · Training & workshops · HR & recruiting by Crosby

Defining the frontier of AI in law through elite research and mentorship.

A six-month research fellowship for PhD candidates to solve the hardest problems in legal AI alongside elite attorneys and engineers.

Launched
2026-06-17

Category: Legal Services

Duration
6 months
Level
PhD candidates and researchers
Team
Crosby's practicing attorneys and engineering team
Program Duration
6 months
Target Audience
PhD Candidates and Researchers
Primary Research Areas
Verification, Personalization, Strategy
  • 6-month research program for PhD candidates
  • Access to unique, non-client contract data sets
  • Regular sessions with elite practicing attorneys
  • Focus on AI verification and strategic negotiation

Overview

The Crosby Intelligence Fellowship is a six-month research program designed for PhD candidates and researchers to tackle the most complex problems in legal AI. Participants work alongside Crosby's engineering and legal teams to develop methods for verifying AI outputs, modeling strategic negotiation, and building personalized review assistants. The program provides fellows with access to unique, real-world non-client contract data and regular mentorship sessions with elite practicing attorneys. Fellows are encouraged to publish their findings to help define the frontier of how AI can sharpen human legal judgment. The inaugural 2026 class includes researchers from Stanford and Princeton focusing on reliability, human-agent collaboration, and strategic decision-making.

Research Focus Areas

  • AI Verification Research: Fellows conduct research into verifying what AI systems produce, such as developing methods for understanding large language model outputs including citations and data attribution.
  • Strategic Decision Modeling: Participants explore how AI makes strategic decisions in complex legal settings, focusing on multi-issue negotiation and litigation strategies.
  • Personalized Review Assistants: Research projects include building review assistants that learn and adapt to the specific priorities and interaction needs of individual human reviewers.
  • Benchmarking and Evaluation: Fellows work on designing and collecting data for RedlineBench, a benchmark that uses preference data to measure AI performance in contract negotiation.

Program Benefits

  • Attorney Mentorship: Fellows have regular sessions with Crosby's practicing attorneys to understand real-world legal judgment and expert feedback.
  • Data Access: Participants are granted access to Crosby's non-client contract data to ground their research in realistic legal scenarios.
  • Publication Freedom: The program provides fellows with the freedom to publish their research findings to help define the frontier of legal AI.
  • Collaborative Environment: Fellows work side-by-side with elite lawyers and engineers in a talent-dense environment to bridge the gap between research and practice.

What's included

  • Research projects, attorney mentorship, data access, and publication rights

Deliverables

  • Research findings, benchmark data, and published papers

Use cases

  • PhD candidates conduct research on verifying large language model outputs using real-world legal data.
  • Researchers build AI assistants that learn to adapt to the specific priorities of human legal reviewers.
  • Fellows develop benchmarks like RedlineBench to measure how AI systems perform in live contract negotiations.
  • Academic researchers collaborate with practicing attorneys to model strategic decisions in multi-issue litigation.

Specifications

Duration
6 months
Level
PhD candidates and researchers
Delivery
Hybrid (in-person sessions with attorneys and research teams)
Team
Crosby's practicing attorneys and engineering team
Availability
Open to researchers from institutions such as Stanford, Princeton, and Columbia Law School
Program Duration
6 months
Target Audience
PhD Candidates and Researchers
Primary Research Areas
Verification, Personalization, Strategy
Mentorship Source
Practicing Attorneys and Engineers
Data Access Type
Non-client contract data
Output Type
Published research and benchmarks

How it works

  1. Application and Selection: Researchers apply to work on open problems in legal AI, such as measuring quality or building review assistants.
  2. Research and Mentorship: Fellows are paired with Crosby's practicing attorneys and engineering teams for a six-month research period.
  3. Development and Testing: Participants use Crosby's data and expert feedback to develop new models, benchmarks, or verification methods.
  4. Publication and Contribution: Fellows publish their findings and contribute to industry standards like RedlineBench.

FAQ

How long is the Crosby Intelligence Fellowship?

The fellowship is a six-month program where researchers work on the hardest open problems in legal AI.

What kind of data and mentorship do fellows receive?

Fellows have access to Crosby's non-client contract data and regular sessions with the firm's practicing attorneys.

Who is eligible for the fellowship?

The program is primarily designed for Computer Science Ph.D. candidates and researchers from leading universities.

Can fellows publish their research?

Fellows have the freedom to publish their research findings during the course of the program.

Added by AntSpace from public sources (intelligence.crosby.ai).

More services by Crosby