CIETAC issues AI guidelines in APAC first

The China International Economic and Trade Arbitration Commission published provisional guidelines on 18 July, setting out high-level guidance for parties, tribunals and other arbitration participants on how to approach the use of AI in international arbitration proceedings. The move, in line with the country’s overarching aim to become a world leader in AI by 2030, makes CIETAC a forerunner in the region for advancing AI use in arbitration.

The guidelines attempt to recognise and support the procedural efficiency benefits of AI tools, while identifying necessary safeguards to manage any potential risks – including data breaches or unauthorised disclosure of sensitive data – that could arise during proceedings and hearings.

CIETAC’s guidelines also outline how AI can be used to support, but must not replace, human-decision making, highlighting the potential for AI-generated outputs to contain inaccuracies or biases that could risk undermining the fairness of proceedings. They stress that arbitrators must retain full responsibility for the reasoning and decision-making set out in awards and that they must ensure that the parties’ rights to be heard are preserved.

They also offer practical recommendations as to what AI-related clauses are permissible in arbitration agreements. In turn, the guidelines caution parties that the use of AI tools does not relieve the parties of their responsibility to guarantee the authenticity and legality of evidence and arbitration-related documents that they submit.

The guidelines also provide a checklist for reducing the risks of using AI in arbitration, including recommending that proper provision is made in any procedural orders or during pre-hearing conferences to regulate such matters.

The guidelines come as courts across APAC, including China, the Hong Kong Special Administrative Region (SAR), Australia, New Zealand and Singapore, also move to respond to the increasing integration of AI into legal proceedings. In November 2024, Singapore published guidance which specifically targeted the use of generative AI (genAI) in court proceedings, requiring all court participants who use genAI to comply.

“The guidelines build on CIETAC’s progressive recognition of the impact of digitalisation and technology set out in its 2024 arbitration rules,” said Rachel Turner, an arbitration expert at Pinsent Mason. “As the first arbitral institution in APAC to develop discussion into guidelines around AI, CIETAC has signaled its intent to remain at the forefront of new developments in arbitration as required by its users.”

Turner added that CIETAC’s initiative to provide practical and helpful guidelines would encourage parties and tribunal members to “engage with the benefits and risks associated the increasing use of AI in arbitration proceedings.”

CIETAC’s guidelines mark the latest efforts by international arbitral institutions to offer guidance and risk mitigation measures for AI use in their proceedings.

In April, the Chartered Institute of Arbitrators (CIArb) issued its own guidelines on regulating the use of AI in international arbitration. “Where CIETAC’s approach is more operational, offering recommendations for procedural orders and party agreements, CIArb’s guidelines are more principle-driven and expansive,” said Mohammed Talib, an expert in international arbitration at Pinsent Masons.

CIArb provides template clauses and procedural orders and outlines arbitrators’ powers to regulate AI use during proceedings. In its guidelines, CIArb also stresses that the arbitral tribunal is ultimately responsible for drafting the award and that it "shall not transfer its decision-making power to AI".