The Use of Artificial Intelligence in Arbitration

Artificial intelligence (AI) is actively making its way into international arbitration, promising to speed up work and reduce costs. However, a high-profile lawsuit by a gamer against Valve and expert opinions show that without clear rules, the use of technology could undermine the foundations of trust in this institution.

1. What is happening?
Artificial Intelligence (AI) is no longer a futuristic concept but a reality in international arbitration. Technologies are being used to analyze documents, predict case outcomes, and even select arbitrators. However, this process is accompanied by serious debates about where assistance ends and the substitution of human judgment begins.

2. The Two Sides of AI: Helper or Threat?

• Benefits:
  • Accelerated data processing and reduced costs.
  • Increased accuracy in case outcome predictions.
  • Rapid document analysis and precedent search.
• Risks:
  • Undermining key principles: human judgment, confidentiality, and trust.
  • The risk of a “mechanized substitution” of justice.

3. Implementation Examples: Three Key Platforms

• Jus Mundi: Uses the Jus AI assistant for multilingual legal research, document analysis, and access to unique archives.
• Dispute Resolution Data: Focuses on predictive analytics, visualizing data on case outcomes and their financial aspects.
• Arbitrator Intelligence: Specializes in collecting data on arbitrators through questionnaires to make their selection more transparent and justified.

4. The High-Profile Case: LaPaglia vs. Valve Corp.
An incident that became a turning point in the AI discussion:

• The Gist: Gamer John LaPaglia filed a petition to vacate the arbitral award, claiming the arbitrator used ChatGPT to draft it.
• Arguments of the Parties: The claimant found uncharacteristic phrasing and unsubstantiated information in the decision. Valve stated that AI was used only for text structuring, while the logic of the decision remained with the arbitrator.
• Significance: A US court is now deciding whether such use of AI constitutes a violation. Experts believe the ruling in this case will set future standards.

5. Expert Opinion: Where to Draw the Line?

• Barry Appleton (Center for International Law, New York Law School): Points to a violation of the intuitu personae principle (trust in personal judgment). “If an arbitrator is unable to independently explain the line of reasoning formed by AI, then there is no arbitration to speak of.”
• Valeria Butyrina (Russian Arbitration Center): Believes that AI can currently only be used for tasks not related to a specific dispute (e.g., comparing rules). Confidentiality and the quality of the process must take precedence over the desire to optimize costs.
• Jonathan Fitch (SVAMC): Predicts that AI will make document analysis and predictive analytics more accessible, but ethical obligations and standards will be tightened.

6. How is Regulation Beginning?
Arbitration institutions are already issuing the first guiding principles:

• Silicon Valley Arbitration and Mediation Center (SVAMC): Prohibits delegating decision-making to AI and requires disclosure of its use.
• Chartered Institute of Arbitrators (CIArb): Defines AI as an auxiliary tool and places full responsibility for the final decision on the arbitrator.

7. The Bottom Line
The future of arbitration lies in a hybrid model where AI handles routine tasks, and humans retain key decision-making. This requires:

1. Transparency: Disclosure of AI use.
2. Control: A ban on delegating final decision-making to algorithms.
3. Security: Guarantees of confidentiality for the parties’ data.

Only then can the technology enhance efficiency without destroying the fundamental principles of arbitral justice.