CIArb Guideline: Can you use AI in arbitration?

Following our recent Keynote on the use of Artificial Intelligence (AI) in court cases, our attention switches to AI in arbitration. The Chartered Institute of Arbitrators (CIArb) recently produced a Guideline on the use of AI for participants in arbitral proceedings and there is a lot of interest in how useful AI could be. Our technology lawyer James Tumbridge and paralegal Anne-Marie Harding share their key take aways on the Guideline.

Part I – Benefits and risks of the use of AI in arbitration

AI is a useful tool for certain tasks, and the CIArb agrees, particularly with regard to legal research, data analysis, and document drafting and review. However, a person needs to be responsible for the use and outputs, so it needs use by someone willing to stand behind the outputs, lest we have embarrassing made-up case law submissions. Well-used AI can deliver great savings, like transcription and translation services, provided all is done in a supervised manner.

Of course, the use of AI does not come without its risks and the CIArb Guideline highlights the following areas that will require consideration, depending on the nature of the arbitration and the AI tool being used:

• Confidentiality: Using a third-party AI tool in arbitration poses significant confidentiality risks and may jeopardise the privacy of sensitive information if data security measures are inadequate. It is crucial to assess each tool’s privacy protections, obtain appropriate permissions, and consider the use of confidentiality agreements to safeguard data.
• Data integrity and security: Following on from the previous point and, given the increased use of electronic, as opposed to hard-copy, submissions, cybersecurity is of vital importance. The CIArb cautions participants in arbitral proceedings to remain vigilant of any enhanced cybersecurity threats arising from the use of AI.
• Impartiality and independence: Arbitrators should take steps to limit the risks of potential algorithmic bias which can arise when using certain AI tools and should be prepared to assume responsibility for the outcome of their decisions.
• Due process: The use of AI tools in arbitration may create due process concerns, such as unequal access to technology and limited consideration of arguments if decisions are influenced by AI-selected data.
• The “black box” problem: The “black box” nature of AI tools (in that it is not often transparent as to how they arrive at a particular outcome) is a concern, and the CIArb recommends a cautious approach when using information generated by an AI tool.
• Enforceability of arbitral awards: Due to the rapid advancement in AI technology and the related legislative and regulatory initiatives, parties must ensure their choice of AI tool does not conflict with any Mandatory Rule, applicable laws, regulations or policies, or institutional rules.

Part II – General recommendations

Participants are encouraged to ensure that they take steps to understand the functionality of any potential AI tool and to weigh up any perceived benefits against the risks and other adverse implications. Parties should also make sure that they are aware of any relevant AI-related law, regulation, or rule of court.

It is important to note that the use of an AI tool does not absolve any participant of its responsibility and accountability.

Part III – Parties’ use of AI in arbitration

The CIArb Guideline sets out the responsibilities of each participant when it comes to the use of AI in arbitration proceedings.

Arbitrators have the overall authority to regulate (except in the case of private use of AI, which is generally permitted if it does not interfere with the integrity of the arbitration), give directions and make procedural rulings on the use of AI, providing that it aligns with applicable laws and regulations or agreements between the parties. AI experts may also be appointed to assist with understanding the implications of the use of an AI tool. Any decision on the use of AI should be documented so that parties can revisit this decision if necessary, during the course of the arbitration.

The parties are permitted to use their autonomy to agree if and how AI is to be used and may propose specific AI tools (again with the caveat that this is in line with applicable laws and regulations). This should also be clearly provided for in the arbitration agreement.

Should the parties disagree on the use of AI, or if the arbitrator considers that the use of AI jeopardises the integrity of the proceedings, the arbitrator can make a ruling on its use. It is important that arbitrators also assess any potential bias in AI tools.

The CIArb Guideline emphasises the importance of transparency. Disclosure of the use of AI tool is of particular importance.

Arbitrators can impose AI-related disclosure obligations on the participants and may make associated directions. Parties should be aware that the failure to comply with directions on the use of AI may result in adverse consequences and may have costs implications.

Part IV – Use of AI by arbitrators

The CIArb Guideline is clear that, whilst arbitrators should not allow AI to make decisions for them, it can be used as a supportive tool in making the process more efficient.

Arbitrators should never allow AI to compromise the integrity of the proceedings and are cautioned that its misuse could affect the validity of an award. All information obtained through AI should be independently verified to ensure accuracy. Arbitrators should also ensure that they are transparent with the parties regarding the use of any AI tool.

If you have questions or concerns about the use of AI in arbitrations, or AI in general, please contact James Tumbridge.