PDCP sets guardrails for personal data in GenAI development

Consumers will receive notifications seeking specific consent in GenAI model training.

Singapore's Personal Data Protection Commission (PDPC) issued new guidelines on the use of personal data in generative artificial intelligence (GenAI), clarifying how data can be collected and used to develop GenAI models.

It also allocates data protection responsibilities across the AI lifecycle and addresses new data sources such as end-user prompts.

Under the new rules, consumers will receive notifications seeking specific consent in certain cases of Gen AI model training.

The Infocomm Media and Development Authority (IMDA) also released complementary Transparency Guidelines for Gen AI Chatbots.

It encourages chatbot deployers to provide "Chatbot Info Cards" covering a chatbot's capabilities and limitations, safety and reliability practices, data use and protection, and reporting channels.

Separately, PDPC updated its Guide to Data Protection Practices for ICT Systems to reflect current best practices drawn from recent data breaches and cyber-essential measures.

On Digital Twin technology, IMDA launched the Digital Twin for Enterprises Playbook, described as the first practical field guide developed with local business context, drawing on input from early adopter enterprises, solution providers and technology leaders in Singapore.

Moreover, on Privacy Enhancing Technologies (PETs), PDPC released a new Guide on Federated Learning developed in consultation with industry experts including Nvidia, and updated its Guide on Synthetic Data Generation to reflect current use cases and practices.

IMDA's PET Sandbox will expand this year with grants for PET proof-of-concept projects and new demo access to PET tools, alongside new sandbox use cases published with Singapore General Hospital and Ant International.