Caution on AI in arbitration

Daily Express

28-07-2025

Published on: Monday, July 28, 2025
Published on: Mon, Jul 28, 2025
By: Sisca Humphrey Text Size: Moderator and panelists during the talk session. Kota Kinabalu: Legal experts are urging the arbitration community to tread carefully as Generative Artificial Intelligence (AI) tools become increasingly integrated into dispute resolution processes. The warnings were delivered during the final panel session at the Bicam Global ADR Horizon 2025 event, themed 'Opportunities and Risks of Generative AI in Arbitration.' The discussion, moderated by High Court of Malaya (Commercial Division) Judge Justice Atan Mustaffa Yussof Ahmad, brought together senior practitioners from across Asia to examine both the practical applications and legal grey areas surrounding the use of AI in arbitration. 'An arbitrator's responsibility is to apply judgment, not to rely on auto-generated text. 'We can benefit from technology, but we cannot delegate our core function,' he said. Senior Partner at EPLegal Tony Nguyen emphasised on how generative AI is already being used to streamline legal workflows, including drafting, document review and issue spotting. 'There's no doubt that these tools improve efficiency, especially for large-scale commercial disputes. 'But speed cannot come at the expense of accuracy or legal coherence,' he said. Tony cautioned against uncritical use of AI-generated content in formal submissions, stressing that most tools are trained on datasets that do not reflect regional legal practices or current arbitration rules. 'We need to ask, is this tailored to the tribunal before us or is it just generic output that sounds plausible?' he said. Head of International Litigation and Arbitration at Skadden Asia Friven Yeoh warned of the risks of factual errors and fabricated references often produced by generative systems. 'What concerns me most is not the occasional mistake. It's that these tools present their results with a tone of certainty, which can mislead even experienced users,' Friven said. He stressed the need for legal practitioners to retain full accountability for any materials produced with AI assistance. 'If you submit it, you're responsible for it. That doesn't change just because you used a machine to help draft it,' he said. He also urged the arbitration community to develop internal guidelines rather than wait for regulation to be imposed externally. Dispute Resolution Lawyer from Singapore Chew Kei-Jin highlighted the limitations of AI in assessing human factors which is a key element in many arbitrations. 'No machine can interpret tone, facial expressions or a witness's hesitation during cross-examination. 'These things influence credibility and outcomes. That's not something AI can replicate,' he said. He acknowledged that AI may be useful in certain backend functions, such as document translation and legal research, but insisted these benefits should not be confused with legal judgment. 'It's a tool, not an advisor, not an advocate and certainly not a decision-maker,' he said. Partner at Kim & Chang in Seoul Matthew Christensen raised concerns over a lack of consistent guidance across jurisdictions. 'Some courts are asking whether AI was used to draft legal documents. Others haven't even considered the question,' he said. He maintained that in arbitration, where confidentiality is crucial that even limited use of AI warrants full disclosure, especially when sensitive data is involved. 'If arbitrators or counsel use AI, parties deserve to know. That transparency matters. 'And we should be careful not to expose client data to third-party systems without clear safeguards,' he said. The panel closed with a reminder from Atan that Malaysia and the wider Southeast Asian region, must engage in structured discussions to clarify expectations around the use of AI in legal practice. 'This is not about rejecting technology. It's about developing a responsible approach to its use in proceedings that affect rights, reputations and businesses,' he said. While no clear consensus was reached on how to regulate AI in arbitration, all speakers agreed that the conversation must continue and that human oversight will remain essential for the foreseeable future. * Follow us on our official WhatsApp channel and Telegram for breaking news alerts and key updates! * Do you have access to the Daily Express e-paper and online exclusive news? Check out subscription plans available.
Stay up-to-date by following Daily Express's Telegram channel. Daily Express Malaysia