Singapore - Generative AI Use in Courts (Registrar's Circular No. 1/2024)

Guide on the Use of Generative Artificial Intelligence Tools by Court Users (Registrar's Circular / Singapore Courts)

Singapore

RAI-SG-NA-GUGAIXX-2024
Effective: October 1, 2024
In Force(In Force)
RegulationTransparency and DisclosureData Protection and PrivacyAccountability and Documentation
Export PDF

The Singapore Courts' Guide (Registrar's Circular No. 1 of 2024) provides principles and practical guidance on the use of generative AI (GenAI) by court users across the Supreme Court, State Courts and Family Justice Courts effective 1 October 2024. It neither bans GenAI nor removes existing professional, evidential or confidentiality obligations — users remain fully responsible for content, must verify AI outputs and must not use GenAI to fabricate evidence.

Summary

The Guide on the Use of Generative Artificial Intelligence Tools by Court Users (Registrar's Circular No. 1 of 2024), issued by the Registrar of the Supreme Court and dated 23 September 2024, applies from 1 October 2024 to all matters in the Supreme Court, the State Courts (including Small Claims, Employment Claims and Community Disputes Resolution Tribunals) and the Family Justice Courts. The Guide defines Generative AI broadly as software that generates content (text, images, audio or video) based on user prompts, and distinguishes such tools from simple spelling/grammar correctors. It recognises that Generative AI can produce coherent, human-like outputs but emphasises that these outputs are not guaranteed to be factually accurate and can hallucinate, fabricate cases or misattribute sources. The Court therefore adopts a neutral position: Generative AI use is not prohibited provided the Guide's requirements are met.

Core obligations set out by the Guide require Court Users — including lawyers, prosecutors, self-represented litigants and witnesses — to remain fully responsible for all Court Documents. Users must fact-check, proof-read, adapt and verify AI-generated content; confirm that any cited case law, legislation or sources actually exist and support the propositions attributed to them; and ensure compliance with intellectual property rules, legal privilege, confidentiality orders and personal data protection laws. The Guide specifically forbids using Generative AI to create, fabricate, embellish or otherwise tamper with evidence relied upon in Court. While a first-draft affidavit or statement may be generated by GenAI subject to the Guide, evidentiary material must not be the product of fabrication.

Operationally, the Guide instructs that the Court may ask a user (and require an affidavit) to declare whether AI was used and to demonstrate compliance if there are grounds to believe AI-generated content was deployed. Users must be prepared to identify portions of their documents that relied on AI and explain how verification was performed. The Guide points users to authoritative verification sources (e.g., Singapore Courts judgments pages and Singapore Statutes Online) and warns that it is insufficient to rely on one Generative AI tool to verify output from another.

Consequences of non-compliance are aligned with existing procedural and professional frameworks: the Court may order costs (including personal costs against lawyers), disregard or discount documents, take disciplinary action against lawyers, or pursue remedies under intellectual property, personal data or contempt-of-court rules. The Guide reiterates that nothing in it displaces existing statutes, rules, codes of conduct or practice directions.

This policy is practical and risk-focused rather than prescriptive: it promotes verification, transparency when asked, protection of confidential data, and professional accountability. It forms part of the Courts' broader approach to modernising practice while safeguarding procedural fairness, evidentiary integrity and privacy. The full text of the Guide is published as Registrar's Circular No. 1 of 2024 and is available from the Singapore Judiciary's official website.

Full article

Read full text ↗

Overview

The Guide on the Use of Generative Artificial Intelligence Tools by Court Users (Registrar's Circular No. 1 of 2024) was issued by the Registrar of the Supreme Court on 23 September 2024 and came into effect on 1 October 2024. It applies to the Supreme Court, the State Courts (including Small Claims Tribunals, Employment Claims Tribunals and Community Disputes Resolution Tribunals) and the Family Justice Courts. The Guide adopts a neutral stance towards Generative AI (GenAI): it does not prohibit its use, but frames mandatory obligations and practical safeguards that place responsibility for accuracy, confidentiality and compliance squarely on the Court User. The full text is published by the Singapore Judiciary; see Registrar's Circular No. 1 of 2024 (Supreme Court page) and the downloadable PDF at Registrar's Circular No. 1 2024 (PDF).

Definitions

The Guide defines key terms to remove ambiguity: "Artificial Intelligence" or "AI" is framed as technology capable of tasks that require intelligence (reasoning, problem solving, learning, planning) but excludes tools that merely correct spelling or grammar. "Generative AI" is explicitly described as software that generates content in response to prompts (text, images, video, audio) — typically using Large Language Models (LLMs). "Court Document" is broadly defined to include any material filed or submitted to Court (text, images, sounds, videos, data), including written submissions, pleadings and affidavits. "Court User" covers all participants in proceedings: lawyers, prosecutors, self-represented litigants and witnesses. These definitions underpin the Guide's scope and requirements.

Governance and Institutional Framework

The Guide sits within the judiciary's procedural governance architecture and does not itself create new legislative duties; instead it clarifies how existing professional, evidential and confidentiality rules apply when GenAI is used. Responsibility rests with Court Users and, for legal practitioners, with existing professional conduct obligations. The Registrar's Circular mechanism implements the Guide across courts; it complements Practice Directions and court rules rather than supersedes them. The Judiciary's pages publish the circular and related resources: Supreme Court Registrar's Circulars and court guides at Court Guides (Judiciary).

Key Focus Areas

The Guide emphasises several focus areas: accuracy and verification of AI outputs; protection of confidential, privileged and sensitive information; intellectual property compliance; evidential integrity; and professional accountability. It clarifies that GenAI tools can "hallucinate" (generate fabricated cases/quotes) and that outputs may be incomplete or outdated. Users must verify any AI-provided legal citations against authoritative repositories (the Guide explicitly cites the Singapore Courts judgments pages and Singapore Statutes Online) and cannot use GenAI to create evidence or tamper with evidentiary material. The Guide advises caution about public AI systems that may store prompts and outputs and warns about potential breaches of confidentiality and personal data protection laws. It also requires users to be able to identify AI-generated portions and to explain verification steps if the Court requests them.

Implementation Framework

Implementation is operationalised through the Registrar's Circular. The Guide becomes binding as court administrative guidance with the authority to inform orders and directions in proceedings. Court Users may use GenAI for drafting or research but must (i) assess suitability, (ii) fact-check and edit outputs, (iii) verify references against authoritative sources, and (iv) ensure compliance with IP, privilege and PDPA/confidentiality requirements. Where the Court has grounds to believe GenAI was used, a user may be asked to inform the Court, declare compliance with the Guide and may be required to file an affidavit. The Guide does not require pre-emptive disclosure of GenAI use unless the Court specifically asks.

Monitoring and Evaluation

The Guide provides for reactive monitoring by the Court in the course of proceedings: judges or registrars may require explanations, declarations or affidavits regarding AI use. The Court may assess whether documents comply with the Guide when considering costs, evidentiary weight or making procedural orders. The Guide also expects Court Users to self-monitor (e.g., using authoritative legal databases for verification) and to adopt practices that minimise risk (secure, enterprise-grade tools where confidentiality is required). As technology and legal practice evolve, the Judiciary can update Registrar's Circulars and Practice Directions to reflect lessons learned.

Penalties, Liability, and Appeals

Enforcement leverages existing judicial and professional mechanisms: non-compliance may lead to orders for costs (including personal costs against counsel), disregard or reduced weight of documents, disciplinary action by the Law Society or other professional regulators, and actions under IP, personal data protection or contempt laws. The Guide explicitly lists potential consequences and ties liability to the responsible Court User (or lawyer) who filed the material. Appeals and remedies remain governed by statutory routes and existing practice directions; any contest about the Guide's application would proceed through ordinary appellate or disciplinary processes.

Relationship to Other Instruments

The Guide clarifies that it does not supersede statutes, rules of court, codes of conduct or existing Practice Directions. It cross-references and relies on existing authoritative legal sources — for verification this includes the Judiciary's judgments portal and Singapore Statutes Online. It complements sectoral and national AI governance initiatives but remains a court-specific administrative instrument that interfaces with professional regulation, data protection (PDPA), intellectual property law and contempt procedure.

International Alignment

While tailored to Singapore's courts, the Guide embodies widely-shared judicial concerns found in other common-law jurisdictions: the need for verification of AI outputs, protection of sensitive data, prohibition on fabricated evidence, and professional accountability. The Guide's neutral, risk-based posture aligns with international judicial approaches that permit tool use subject to verification and disclosure policies. It therefore supports interoperability with multinational practice (e.g., for cross-border litigation or the Singapore International Commercial Court) while preserving local procedural safeguards.

Implementation Timeline

EventDate
Registrar's Circular issued (signed)2024-09-23
Registrar's Circular published online2024-09-26
Guide effective date2024-10-01

Sources and References

SourceType
Registrar's Circular No. 1 2024 — Guide on the Use of Generative Artificial Intelligence Tools by Court Users (PDF)Primary Source
Registrar's Circular No. 1 2024 (Supreme Court web page)Primary Source

Requirements for a company

What an organisation has to do under Singapore - Generative AI Use in Courts (Registrar's Circular No. 1/2024), at a glance. Not legal advice — the table below gives the provision and deadline for each item.

Must do

9
  • Do not use GenAI to create or tamper with evidentiary material.All Court Users
  • Verify all AI-generated legal citations against authoritative sources.All Court Users
  • Protect confidential, privileged, and sensitive information from GenAI systems.All Court Users
  • Ensure compliance with intellectual property laws when using GenAI outputs.All Court Users
  • Ensure compliance with personal data protection laws when using GenAI outputs.All Court Users
  • Fact-check and edit all GenAI-generated outputs for accuracy before submission.All Court Users
  • +3 more in the table below

Must not do

0

Nothing in this category.

Should do

1
  • Adopt practices that minimise risks, such as using secure, enterprise-grade GenAI tools for confidential matters.All Court Users

Should not do

0

Nothing in this category.

Who must do what

The obligations under Singapore - Generative AI Use in Courts (Registrar's Circular No. 1/2024), most serious first. Not legal advice — verify against the official text before relying on it.

#WhoRequirementBy whenWhereSeverity
1All Court UsersDo not use GenAI to create or tamper with evidentiary material.
cannot use GenAI to create evidence or tamper with evidentiary material.
Before filing/submitting to CourtCritical
2All Court UsersVerify all AI-generated legal citations against authoritative sources.
Users must verify any AI-provided legal citations against authoritative repositories
Before filing/submitting to CourtCritical
3All Court UsersProtect confidential, privileged, and sensitive information from GenAI systems.
Do not input privileged or sensitive information into public GenAI systems
Before inputting data into GenAICritical
4All Court UsersEnsure compliance with intellectual property laws when using GenAI outputs.
ensure compliance with IP, privilege and PDPA/confidentiality requirements.
Before filing/submitting to CourtCritical
5All Court UsersEnsure compliance with personal data protection laws when using GenAI outputs.
ensure compliance with IP, privilege and PDPA/confidentiality requirements.
Before filing/submitting to CourtCritical
6All Court UsersFact-check and edit all GenAI-generated outputs for accuracy before submission.
fact-check and edit outputs
Before filing/submitting to CourtCritical
7All Court UsersMaintain full responsibility for all content generated or assisted by GenAI.
users remain fully responsible for content
OngoingCritical
8All Court UsersBe prepared to identify AI-generated portions and explain verification steps if requested by the Court.
requires users to be able to identify AI-generated portions and to explain verification steps if the Court requests them.
When requested by CourtImportant
9All Court UsersAssess the suitability of GenAI tools for the specific legal task.
assess suitability
Before using GenAI for a taskImportant
10All Court UsersAdopt practices that minimise risks, such as using secure, enterprise-grade GenAI tools for confidential matters.
adopt practices that minimise risk (secure, enterprise-grade tools where confidentiality is required).
OngoingRecommended

© Regulations.AI · updated on 13-Jun-2026