The Chartered Institute of Arbitrators (“<span class="news-text_medium">CIArb</span>”) has introduced its Guideline on the Use of AI in Arbitration (2025) (the “<span class="news-text_medium">Guideline</span>”), designed to support arbitration participants in integrating artificial intelligence (“<span class="news-text_medium">AI</span>”) into their processes while safeguarding the integrity of arbitration proceedings. The Guideline complements applicable laws, regulations and arbitration rules, aiming to help users leverage AI's advantages while minimising risks to due process, fairness and the enforceability of awards.
The Guideline is divided into four sections. The first section explores the advantages and risks of using AI in arbitration, while subsequent sections provide general recommendations and cover the use of AI by both parties and arbitrators. It also includes a template agreement for party use of AI tools, along with two template procedural orders to help arbitral participants incorporate AI into their proceedings effectively.
One of the key principles in the Guideline is party autonomy. It acknowledges that parties have the freedom to agree on whether and how AI tools can be used, subject to mandatory laws or regulations. In cases where the parties disagree on the use of AI, the tribunal can make a decision, considering various factors such as the potential impact on evidence, fairness, equality of arms, confidentiality and the specific characteristics of the AI tools, including their data quality, accuracy, bias and security.
The Guideline outlines the powers of arbitrators to require the disclosure of AI tool use if it may affect evidence or the outcome of the arbitration. Arbitrators are also empowered to appoint AI experts to help them understand AI tools and their implications. It notes that tribunals cannot regulate the “private use of AI by parties”, as long as such use does not interfere with the proceedings or compromise the integrity of the arbitration.
In determining whether AI-assisted or AI-generated content should be included in the record, arbitrators are encouraged to verify that the source data for the machine-generated content is part of the record. They may also request the parties to clarify how the inputs to the AI system are connected to the outputs it produces.
Arbitrators can utilise AI themselves to enhance the accuracy and efficiency of processing information but must independently verify the information obtained through AI tools. Importantly, the use of AI should not replace legal analysis, research, or the interpretation of facts and law, as it could unduly influence procedural or substantive decisions. The guideline encourages arbitrators to consult with the parties and fellow tribunal members before incorporating AI into their decision-making processes.
