AI in arbitration: A new frontier?

Artificial Intelligence (AI), once considered peripheral to arbitration, has experienced a remarkable surge in adoption over the past five years.

The appeal is obvious - the proliferation of AI tools at a lawyer's disposal promises significant reductions in legal costs and substantial gains in efficiency. Nevertheless, arbitral bodies are urging practitioners to approach these tools with caution. This is not only due to the well-documented challenges associated with AI, but also given the current absence of comprehensive regulatory guidance and a limited body of test cases addressing the use of AI in a legal environment.

AI's current state of play in arbitration

The landscape of AI in arbitration is rapidly evolving, and there are several examples of how AI tools are currently being utilised.

One of the most mature applications of AI is in factual and legal research. Advanced platforms are enabling users to scan millions of cases, statues and arbitral awards to identify relevant material and generate helpful summaries. AI tools are also used to analyse case law to highlight areas where a practitioner's argument may be vulnerable to adverse or unsettled precedent, as well as identify and produce authorities with similar language or reasoning, thereby facilitating efficient responses to nuanced legal questions.

AI is also increasingly utilised in case management and procedural logistics, particularly in complex international disputes. AI tools can coordinate party availability and generate procedural timetables that account for flights, commitments and different time zones. Similarly, AI-powered translation tools provide real-time simultaneous translation, enabling arbitrators to manage proceedings accurately in multiple languages, whether in person or virtually.

In the realm of document review and evidence management, AI helps to process and manage large volumes of evidence. By employing sophisticated keyword searches and pattern recognition, AI tools can swiftly locate relevant documents, significantly reducing the time required to review and produce critical documents and thereby streamlining the evidentiary process for all parties involved.

However, despite its numerous advantages, the use of AI must be tempered by a careful consideration of its risks.

Existing regulatory landscape and what constitutes best practice AI-use

Concerns persist around AI use regarding the potential for undetected errors and algorithmic bias, the risk of data breaches and the inadvertent disclosure of confidential information.

In response to these concerns, several arbitral bodies have introduced guidelines to assist parties in integrating AI into arbitral proceedings, while still managing these risks. For example, the Chartered Institute of Arbitrators (CIArb) has released the Guidelines on the Use of AI in Arbitration (2025), which provide a practical framework for the responsible adoption of AI in arbitration. The guidelines recommend that arbitrators be empowered to impose AI-related disclosure obligations on the parties, including directions concerning when and how disclosure of AI may be required. Similar guidance has been issued by the Stockholm Chamber of Commerce's Guide to the Use of AI in Cases Administered under the SCC Rules, and JAMS' Rules Governing Disputes Involving Artificial Intelligence Systems. Each offer their own protocols for the use of AI in cases administered under their respective rules.

LaPaglia: a cautionary case study on the possibility of new avenues for set aside

The recent case of LaPaglia v. Valve Corp illustrates the potential for AI to give rise to grounds for challenging arbitral awards. In this matter, the claimant (Mr. LaPaglia) attempted to vacate an arbitral award on the basis that the arbitrator had 'utsourced his adjudicative role to Artificial Intelligence'. Specifically, the claimant argued that the award included 'telltale signs of AI generation', including citing facts which 'are both untrue and not presented at trial or present in the record.' Notably, the challenge was brought under section 10(a)(4) of the Federal Arbitration Act (US), which permits set aside where an arbitrator has 'exceeded their power' by acting outside the scope of the parties' contractual agreement.

The claimant contended that by using AI, the arbitrator had exceeded his authority by delegating his decision-making function to a machine. While the court is yet to give its decision, the case highlights the potential for parties to challenge awards where AI is perceived to have played a substantive role in the decision-making process.

Key takeaways in using AI in arbitration

Looking ahead, it is incumbent upon practitioners and advisors to address the use of AI proactively through carefully drafted alternative dispute resolution (ADR) clauses. Such clauses should specify the types of AI technologies that may be employed in the dispute resolution process, delineate their intended use, and establish appropriate safeguards to ensure the accuracy and integrity of the proceedings and the resolution reached. Given the rapid evolution of these emerging technologies, it may be some time before consistent and widely accepted use cases emerge. Nevertheless, the guidance issued by leading arbitral institutions such as CIArb, the Stockholm Chamber of Commerce and JAMS, offers parties a measure of confidence in adopting AI, while also signalling the direction in which the regulatory landscape space is evolving to address the challenges and opportunities presented by these emerging tools.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.

Lawyers Weekly Law firm of the year 2021	Employer of Choice for Gender Equality (WGEA)