Singapore consults on personal data rules for generative AI

Singapore’s Personal Data Protection Commission (PDPC) has launched a public consultation on proposed advisory guidelines governing the use of personal data in generative AI systems. Published on 2 June, the draft guidelines seek feedback on how Singapore’s Personal Data Protection Act (PDPA) applies when personal data is used in the development and deployment of generative AI systems.

The proposed guidelines address the collection and use of personal data for generative AI model development, the allocation of data protection responsibilities across the AI lifecycle, and the handling of individual rights requests relating to personal data. The guidance is organised around development, deployment, and post-deployment stages.

For model development, the draft guidelines clarify how organisations may rely on exemptions for publicly available information when using web-scraped datasets containing personal data. They also set out considerations for data behind digital barriers such as paywalls, registration requirements, authentication mechanisms, and tools that block automated access.

The PDPC proposes that general privacy notices should not be considered sufficient for obtaining consent to use personal data for large-scale AI training or fine-tuning. Organisations would instead be expected to provide AI-specific notices explaining the categories of personal data used, the purpose of the processing, the model’s intended functions, and how individuals can refuse or withdraw consent.

The proposed guidelines also outline responsibilities for model providers, system providers, and system deployers, including retention, protection, purpose limitation, and accountability obligations. The post-deployment guidance addresses access and correction requests while recognising technical challenges associated with large datasets, embeddings, temporary context windows and the removal of specific information from trained models. Interested parties may submit comments to the PDPC by 1 July 2026.

Why does it matter?

The consultation highlights the growing challenge of applying existing data protection laws to generative AI systems that rely on large-scale data collection and model training. Regulators worldwide are increasingly examining how privacy principles such as consent, transparency and purpose limitation should operate in AI development.

Singapore’s proposed guidance could provide an important reference point for organisations developing or deploying generative AI, particularly in areas such as web scraping, AI training datasets and the allocation of responsibilities across the AI value chain.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!