Singapore Legal Sector GenAI Guide

Overview

Singapore’s Ministry of Law launched the Guide for Using Generative AI in the Legal Sector on 6 March 2026 at The Next Charter: Shaping Singapore’s Legal Future Together. The Guide supports safe and responsible use of generative artificial intelligence, or GenAI, in legal services and forms part of the Ministry’s broader efforts to advance legal-sector transformation and digitalisation. It combines principles for responsible use with practical implementation guidance and case studies. The Ministry’s launch announcement describes its purpose, development, and principal themes.

The Guide is a final, officially published, non-binding guideline. Its adopted status identifies publication of the final instrument rather than commencement of legislation. It does not create a separate legislative enforcement regime. Its intended audience includes anyone handling legal work in Singapore, including legal professionals, law practices, smaller practices, and in-house legal teams. The official Guide addresses adoption alongside professional safeguards: use of technology does not transfer responsibility for legal work away from legal professionals.

Terminology and Core Principles

GenAI means generative artificial intelligence, the technology addressed by the Guide. Legal research, enhancing drafting quality, and strengthening knowledge management are identified applications in legal practice. These examples illustrate ways the technology can support legal services rather than defining an exhaustive catalogue of legal tasks. The Guide’s practical orientation connects the use of tools to the responsibilities of the people conducting legal work and to the needs of the organisations and clients they serve.

The three central principles are Professional Ethics, Confidentiality, and Transparency. Professional ethics concerns responsibility for work products and the exercise of requisite knowledge, skill, and professional judgment. Confidentiality concerns safeguarding client information and considering how different GenAI models process and store data. Transparency concerns whether clients should be informed about GenAI use, particularly when that use substantially contributes to legal work, affects service costs, or involves data-handling practices that may not align with client requirements. Together, these principles provide the organising framework for responsible adoption.

Governance and Stakeholder Participation

The Ministry of Law developed the Guide in close consultation with legal-sector stakeholders. Public consultation ran from 1 to 30 September 2025, and the final Guide was refined following feedback from over 20 local and international stakeholders. Participants included law practices, in-house counsel, legaltech providers and consultants, academia, and industry bodies. This development process brought together professional, organisational, technical, and academic perspectives on the use of GenAI in legal work.

The Guide addresses real-world practice needs and diverse operating contexts, including those of smaller law practices and in-house legal teams. Its governance emphasis remains the responsibility of legal professionals for their work. As an implementation approach, an organisation can allocate responsibility for tool selection, client-information decisions, review of outputs, and client communications. Such allocations are practical ways to apply the principles, rather than a prescribed committee structure or staffing model. The Ministry also maintains a continuing role through stakeholder engagement and updates to the Guide as necessary.

Professional Ethics and Accountability

Legal professionals remain ultimately responsible for all work products and must exercise the requisite knowledge, skill, and professional judgment. GenAI assistance therefore does not shift responsibility for legal work to a model or its provider. This principle applies to the professional assessment of GenAI-assisted work as well as the decision to use a tool. A generated answer, draft, or knowledge-management output must be considered within the professional’s responsibility for the work product and its use in the matter.

Legal research, drafting, and knowledge management can benefit from GenAI while still requiring professional assessment. As practical implementation measures, an organisation can identify who reviews assisted work, determine whether the selected tool is suitable for the task, and provide training on the continuing role of professional judgment. These measures translate the accountability principle into working arrangements; they are not a separate statutory duty list. The Guide’s non-binding character does not remove the professional responsibilities it emphasises, and GenAI use should not be treated as a substitute for the knowledge and skill needed to perform legal work.

Confidentiality and Client Information

Users must safeguard client data and confidentiality and remain mindful of the risks associated with different GenAI models. A central consideration is how information is processed and stored. Selecting a tool for legal work therefore involves attention to its data-handling practices as well as its usefulness for the task. The confidentiality principle connects the information entered into a model with the obligations and expectations surrounding client information.

As a practical application of this principle, an organisation can identify the sensitivity of proposed inputs, assess the selected model’s processing and storage arrangements, and compare those arrangements with relevant client requirements. Documenting those decisions can support consistent handling of client information and make responsibilities clear within a team. Such documentation is an implementation suggestion rather than a prescribed format. When a tool or its service arrangements change, an organisation can revisit whether the handling of client data remains appropriate. These considerations are relevant to law practices and in-house teams of different sizes.

Transparency and Client Communication

Legal professionals should consider informing clients about GenAI use, particularly where it substantially contributes to producing legal work, affects the cost of legal services, or involves data-handling practices that may not align with client requirements. The principle calls for consideration of the significance and consequences of the use in the relevant matter. It connects client communication to the contribution of the tool, the economics of the service, and the treatment of client information.

An implementation approach can include assessing these three circumstances when a tool is introduced or when its role in a matter changes. For example, an organisation can consider whether the extent of GenAI assistance has become substantial, whether its use changes service costs, and whether the selected model’s processing or storage arrangements fit the client’s expectations. These are practical decision points derived from the transparency principle, not a universal disclosure script. Client communication can be considered alongside tool selection and confidentiality decisions so that adoption arrangements remain consistent with the client’s requirements.

Practical Implementation and Case Studies

The Guide provides practical guidance on implementing GenAI tools in legal practice and includes case studies showing safe and effective deployment while upholding professional standards. The applications highlighted include legal research, improving drafting quality, and strengthening knowledge management. Examples from organisations of varying sizes support an approach that considers both the benefits of adoption and the safeguards needed in different legal-service settings.

A practical implementation sequence can begin with identifying the legal task, assessing the tool’s suitability, considering the client information involved, and assigning responsibility for reviewing the resulting work. Organisations can then consider client communication, explain the continuing role of professional judgment through training, and record decisions about acceptable inputs and data handling. These steps are suggested ways to operationalise the three principles rather than a prescribed sequence. Smaller practices and in-house teams can apply the same considerations within their own operating arrangements, keeping the focus on responsible use rather than on a particular technology product or organisational model.

Related Guidance and International Context

The Singapore Judiciary’s Guide on the Use of Generative Artificial Intelligence Tools by Court Users applies from 1 October 2024 and is a separate companion instrument for court-related work. The Ministry Guide addresses legal-sector use, while the Judiciary document addresses court users. Organisations undertaking court work can consult the Judiciary guidance alongside the Ministry Guide and applicable procedural instruments. The Infocomm Media Development Authority’s artificial-intelligence page provides broader government context on AI and sector-specific guidance.

The Ministry Guide’s development was internationally informed through participation by both local and international stakeholders. For organisations serving international clients, its confidentiality and transparency principles offer practical considerations concerning model processing, storage, and alignment with client requirements. An organisation can apply those considerations when evaluating tools for matters involving cross-border client needs. The instrument remains Singapore legal-sector guidance, and such practical applications are distinct from authorisation of international data transfers or compliance with foreign law.

Continuing Review and Legal Responsibility

The Ministry of Law will continue engaging stakeholders and update the Guide as necessary to reflect technological developments, evolving best practices, and the legal sector’s needs. This continuing-review approach supports guidance that can respond to changing tools and operating conditions. Organisations can similarly consider reviewing their adoption arrangements when a provider changes its service, a team expands the tasks assigned to GenAI, or client requirements change.

As an implementation suggestion, periodic review can ask whether professionals meaningfully assess work products, whether model processing and storage remain suitable for client information, and whether the scale or consequences of use warrant reconsidering client communication. The Guide is non-binding guidance rather than legislation establishing a separate penalty or appeals regime. Legal professionals’ ultimate responsibility remains central. Professional, procedural, contractual, and other legal obligations applicable to a matter remain separate considerations when assessing liability or compliance; the Guide’s principles should not be treated as immunity from those obligations.

Implementation Timeline

Implementation Checklist

Sources and References