Law / Technology

Add as a preferred source on Google

Singapore Chief Justice Urges Limits on AI in Judicial Decisions

Sundaresh Menon said AI may help organize case information, but judges should remain responsible for analysis and the reasons behind rulings.

Updated 2 min read
Judicial officers attend a conference session in Brisbane, Australia. (AI-generated image / EconoTimes)
Judicial officers attend a conference session in Brisbane, Australia. (AI-generated image / EconoTimes)

Singapore Chief Justice Sundaresh Menon said courts should reserve their greatest caution for AI-generated judicial reasons, arguing that human responsibility remains essential even as tools help manage case information.

Menon spoke at the 28th Annual Australian Judicial Officers Association Colloquium in Brisbane on Oct. 9. He said AI is already entering the justice system through litigants, legal professionals and tools used to administer courts. Growing case complexity, large records and demand that exceeds available judicial resources create pressure to use technology, he said.

Menon described a graduated role for AI. Software may help collect and organize information, while its use in analyzing evidence or suggesting outcomes calls for more caution. He said courts should be most restrained when tools are used to formulate judicial reasons: a polished explanation may not show how a judge reached a decision. “It cannot bear judicial responsibility,” Menon said.

He identified three risks: people may rely too heavily on AI, judges may lose skills through disuse, and system designers, vendors or procurers may gain hidden influence over court processes. Errors or decisions made through opaque systems could affect fairness, judicial independence and public accountability, he said.

Menon urged courts to pilot and audit tools, train judges and design workflows that require judges to engage independently with cases. He also called for a principled standard for deciding when parties should be told AI was used and what that notice should include, along with greater cooperation among judiciaries.

For some high-volume, lower-value disputes, such as unpaid invoices, tenancy deposits and minor motor accidents, courts could use AI to offer an optional preliminary assessment, Menon said. Parties would still be able to take a case to a human decision-maker. He compared the idea with early neutral evaluation in Singapore’s State Courts and technology-assisted processes used by online marketplaces.

Menon also said AI-assisted translation and summaries could help people representing themselves understand case material and navigate court processes. Those potential benefits would need to be weighed against the risks to parties from errors or unfair procedures, he said.

Related articles

View all