Introduction

Arbitrator selection is all about finding and appointing a neutral third party to settle a dispute through arbitration. It’s a key step that affects how fair, quick, and effective the process will be—and even the final decision. Both law firms and arbitration organizations play important roles in picking the right person, each using similar criteria. They look at things like expertise, impartiality, and what’s best for their client.

In this post, we’ll dive into those selection factors and consider whether Deep Research or further enhanced tools down the road can improve arbitrator selection research. We'll also conduct an experiment to see how effectively Deep Research or modified agents can assist with arbitrator selection.

Key Considerations in Arbitrator Selection

Appointing an arbitrator in international arbitration depends on two key factors: I. ratione personae (the arbitrator’s qualities) and II. ratione materiae (the nature of the dispute).

I. Arbitrator-Related Factors (Ratione Personae)

• Eligibility: Arbitrators must be natural persons with legal capacity. Traits like legal/commercial expertise, integrity, and independence are key. Certain roles (e.g., judges in France) may disqualify.

• Training & Reputation: Legal background and good standing are preferred. Some countries (e.g., Argentina) require legal qualifications unless acting as amiable compositeurs. Training and mock arbitrations help.

• Track Record: Past rulings may show bias (e.g., Caratube). Academic vs. firm views, parallel cases, or disqualifications need review.

• Impartiality & Conflicts: Institutions check for neutrality. Red flags include:

  • Pre-appointment contact or hospitality

  • Undisclosed case overlaps (Scandinavian Reinsurance)

  • Repeat appointments or prior counsel disputes

  • Ex parte meetings

• Independence: Requires no ties to parties and full disclosure. It’s distinct from impartiality and comes first.

• Practical Issues: Language, availability, travel, fees, legal system familiarity, nationality, and diversity all matter. Institutions like LCIA and AAA apply specific rules on nationality in cross-border cases.

II. Dispute-Related Factors (Ratione Materiae)

• Nature of the Dispute: Arbitrators need subject-matter expertise. Some institutions (e.g., USSR Chamber of Commerce) require it. AAA, SIAC, HKIAC, and ICSID keep lists of qualified experts. Industry-specific knowledge improves technical and commercial judgment.

• Other Factors: Arbitrator selection may also depend on net worth, business size, transaction type, dispute value, and location.

Current Manual Selection Process

Selecting an arbitrator can take anywhere from several days to weeks, depending on the complexity of the case and the availability of suitable candidates. The process involves thorough background checks, conflict-of-interest assessments, and comparative evaluations to ensure the most qualified and impartial arbitrator is chosen. A typical arbitrator selection process may include:

• Database Searches: Reviewing arbitrator rosters and specialized databases (e.g., Jus Connect, Rising Arbitrators).

• Online Research: Conducting Google searches for reputation insights and past case involvement.

• Professional Networking Sites: Checking LinkedIn for background, endorsements, and affiliations.

• Publication Reviews: Examining legal publications, academic papers, and past rulings to assess expertise.

• Conflict Checks: Cross-referencing arbitration history, advisory roles, and affiliations to avoid conflicts of interest.

This manual process is time-consuming and labor-intensive, often requiring multiple team members. Can AI help and accelerate arbitrator selection?

Potential of OpenAI’s Deep Research

OpenAI’s Deep Research is an AI-powered tool (agent) designed to autonomously conduct comprehensive research by searching, analyzing, and synthesizing information from multiple sources on the web. It breaks down the task into smaller sub-tasks, navigates the internet, looks up multiple queries, and pieces together findings much like a human researcher. While Perplexity AI offers a similar capability, we’ll be running our experiment with OpenAI’s Deep Research, as it provides a more thorough and structured research approach.

Comparing Deep Research to Traditional Arbitrator Selection

Traditional arbitrator selection involves extensive manual effort, including:

• Database Searches: Scouring institutional arbitrator rosters.

• Online Research: Google and LinkedIn searches for professional history.

• Publication Reviews: Analyzing legal papers and previous rulings.

• Conflict Checks: Reviewing affiliations and prior case involvement.

Deep Research can automate arbitrator research by analyzing publicly available data and reducing manual work. However, its effectiveness would vary, as it is not a specialized tool specifically for arbitrator selection but rather a general research AI agent. Used right, tools like Deep Research can significantly improve arbitrator selection.

Experiment Overview

We tested OpenAI's Deep Research to evaluate its ability to recommend suitable arbitrators for international arbitration, specifically focusing on Spanish-speaking experts familiar with ICSID proceedings.

Experiment Setup: We used the Notice of Arbitration (in Spanish) from a recent case, Eléctricas de Medellín Ingeniería y Servicios S.A.S. v. Republic of Honduras (ICSID Case No. ARB/24/24), ensuring the document contained no references to appointed arbitrators or names. The prompt instructed Deep Research to identify arbitrators for respondent based on independence, Spanish proficiency, and expertise relevant to ICSID disputes involving energy and public-private partnerships (PPP).

Sources Checked by Deep Research

Deep Research searched publicly available sources including:

• ICSID's official arbitrator lists and profiles

• Jus Mundi and IAReporter for arbitration case histories

• Italaw for relevant arbitration awards

• Arbitrator professional profiles (Legal500, ArbitrationLaw.com, LinkedIn)

Research Output

Recommended Arbitrators by DeepResearch for Honduras:

• Eduardo Zuleta (Colombia): Extensive ICSID experience, specializes in energy and infrastructure disputes; bilingual (Spanish/English).

• Professor Brigitte Stern (France): Expert in state responsibility and expropriation; multilingual (French, English, Spanish, German).

• Gabriel Bottini (Argentina/Spain): Former Argentine state attorney with deep experience in regulatory disputes and infrastructure projects; native Spanish speaker.

• Dr. Raúl E. Vinuesa (Argentina): Experienced in sovereign-state arbitration and Latin American regulatory law; native Spanish speaker.

• Juan Fernández-Armesto (Spain): Extensive experience in PPP and infrastructure arbitrations; frequently appointed in Latin American cases; native Spanish speaker.

A detailed PDF report of the research output from Deep Research can be accessed here.

Limitations and Future Directions

Deep Research effectively identified arbitrators with relevant language proficiency, regional expertise, and ICSID experience. However, certain important aspects were not fully addressed:

• Availability: The availability or scheduling of arbitrators was not verified.

• Conflict Checks: Basic checks were conducted, but the tool did not identify nuanced or undisclosed conflicts.

• Private Data Sources: The search was limited to publicly available sources, excluding proprietary institutional data.

Conclusion

OpenAI's Deep Research shows promising potential to streamline and enhance arbitrator selection by significantly reducing manual research effort. The experiment demonstrated that the tool effectively identified qualified arbitrators based on language skills, regional expertise, and relevant arbitration experience. However, limitations emerged around conflict-of-interest checks, arbitrator availability, and access to proprietary data—areas critical to the arbitrator selection process.

To maximize AI’s value in arbitrator selection, future developments should integrate specialized conflict-checking capabilities, real-time availability tracking, and secure access to proprietary arbitration databases. While general-purpose AI agents like Deep Research provide a strong foundation, dedicated arbitrator selection tools tailored to the nuances of arbitration practice will deliver the most practical benefits.