Text & data mining in Singapore: Three years on

Introduction

Three years after Singapore enacted its text and data mining (TDM) exception, the copyright framework remains the same, but the surrounding AI governance landscape has grown considerably. Singapore continues to position itself as an AI-friendly jurisdiction, pairing a stable copyright framework with fast-evolving guidance on responsible AI use.

The Singapore Copyright Act 2021 (CA) provides for a fair use exception modeled after the fair use provisions in U.S. copyright law. Under the fair use exception, whether the use of a copyright-protected work qualifies as a non-infringing fair use is assessed according to a number of factors including the purpose and character of the use and the effect on the market for the work.

This fair use exception was nevertheless considered too unpredictable, and a dedicated exception was introduced to give the AI sector more certainty over TDM.

The TDM exception was introduced via sections 243 and 244 of the CA and took effect in November 2021. Three years on, the TDM exception remains unchanged, a sign of regulatory stability even as global approaches to AI training data continue to shift.

Under section 243 of the CA, “computational data analysis,” in relation to a work or a recording of a protected performance, includes:

1.  using a computer program to identify, extract and analyse information or data from the work or recording; and
2.  using the work or recording as an example of a type of information or data to improve the functioning of a computer program in relation to that type of information or data.

This computational data analysis exception extends to communicating the work to the public and publication of the work.

Five conditions under section 244(2) of the CA must be met:

1. the copy must be made only for computational data analysis;
2. it cannot be used for any other purpose;
3. it cannot be shared except to verify results or for related collaborative research;
4. the user must have lawful access to the source material; and
5. where that source is itself an infringing copy, the user must not know (or reasonably suspect) this, unless a narrow exception for necessary uses applies.

Three years on, Singapore’s courts have not decided whether works solely produced by generative AI (as opposed to generative AI assisting human creators) will receive copyright protection.

The likely position remains that such works will not qualify, consistent with the policy aim of discouraging “stockpiling,” where generative AI is used in bad faith to mass-produce works and profit from inflated sales to creators. Overseas, courts and regulators are increasingly confronting disputes over AI training data and the ownership of AI-generated outputs, but no Singapore decision has yet tested these issues locally.

A broader governance layer

Alongside the copyright exception, Singapore has moved to address the rise of agentic AI. On January 22, 2026, the Infocomm Media Development Authority (IMDA) launched the Model AI Governance Framework for Agentic AI, billed as the world’s first framework offering a comprehensive guide for enterprises to deploy agentic AI responsibly. It builds on the original Model AI Governance Framework introduced in 2020 and reflects feedback from both government agencies and private sector organizations.

The Framework guides organizations deploying AI agents, whether built in-house or sourced from third parties, across four dimensions: assessing and bounding risks upfront by choosing appropriate use cases and limiting agents’ autonomy and access to tools and data; making humans meaningfully accountable through defined checkpoints requiring human approval; implementing technical controls and processes throughout the agent lifecycle, including baseline testing and access controls; and enabling end-user responsibility through transparency, education, and training. The IMDA has described the Framework as a living document and continues to invite feedback and case studies from industry for future updates.

Relevant data protection developments

Data protection guidance has also moved to address generative AI directly. The Personal Data Protection Commission (PDPC) issued its final Advisory Guidelines on Use of Personal Data in Generative AI on July 20, 2026. The Guidelines map obligations under the Personal Data Protection Act across the development, deployment, and post-deployment phases of the AI lifecycle. 

The Guidelines clarify when the “publicly available” exception can be relied on to use personal data behind digital barriers such as paywalls or logins, and allocate responsibility across the AI supply chain between model providers, system providers, and system deployers. Businesses training or deploying generative AI in Singapore should now assess both the TDM copyright exception and these updated PDPC expectations.

Looking ahead

Unlike the EU’s opt-out mechanism, Singapore’s TDM exception gives rightsholders no ability to reserve their works from AI training use, which continues to shape licensing negotiations for content owners. Businesses developing or deploying generative AI in Singapore should keep monitoring this fast-moving area, review data licensing and content agreements for TDM-related terms, and ensure their AI governance practices are aligned with the latest official guidance.