Singapore has introduced new data protection guidelines for generative AI alongside a cross-border data agreement with Japan.
The initiatives aim to help businesses adopt AI and data-driven technologies while managing privacy obligations and commercial sensitivities.
Minister for Digital Development and Information Josephine Teo announced the initiatives at the inaugural Singapore Data Festival.
Consent rules for AI training
The Personal Data Protection Commission (PDPC) issued advisory guidelines detailing how data protection principles apply across the AI lifecycle.
Covering areas such as data collection, usage, and accountability, the guidelines clarify how companies can use personal data for Gen AI development. They also recommend specific consent notifications in cases where personal data is used for model training.
Data protection responsibilities across the AI lifecycle, including for new data sources such as end-user prompts, are also outlined in the guidelines.
For consumer applications, the Infocomm Media Development Authority (IMDA) introduced transparency guidelines for generative AI chatbots.
The guidelines encourage deployers to provide Chatbot Info Cards that outline their systems’ capabilities, data use practices, and reporting channels.
IMDA said this will help consumers make more informed decisions about the information they share with AI services.
Lowering cross-border compliance costs
To support businesses operating internationally, PDPC signed a Memorandum of Cooperation with Japan’s Personal Information Protection Commission.
The agreement focuses on promoting the Global Cross-Border Privacy Rules and establishing a basis for developing model contractual clauses.
By supporting these mechanisms, Singapore and Japan aim to reduce compliance costs and friction for businesses transferring data between both markets while maintaining privacy protections.
PDPC Commissioner Denise Wong and Japanese PPC Chairman Dr Tezuka Satoru signed the pact.
Privacy technologies and secure data sharing
PDPC also published a guide on federated learning, developed in consultation with Nvidia, and updated its Guide on Synthetic Data Generation.
These privacy enhancing technologies allow organisations to derive insights from data while reducing the need to share raw datasets.
To support adoption, IMDA will enhance its PET Sandbox by providing access to demonstrations of PET tools and publishing additional use cases.
Recent examples include use cases involving Ant International and Singapore General Hospital.
Alongside these AI and privacy initiatives, IMDA launched a Digital Twin for Enterprises Playbook to help businesses in sectors such as logistics and manufacturing identify use cases and implement digital twin projects.
Featured image credit: Edited by Fintech News Singapore, based on image by magnific via Magnific
