- Singapore has introduced guidelines requiring organisations to inform consumers when their personal data is used to train generative artificial intelligence models, effective July 20.
- The Infocomm Media Development Authority (IMDA) has rolled out new transparency guidelines for generative AI chatbots, aiming to enhance consumer understanding of AI applications.
- Under the new guidelines, organisations must not require individuals to consent to the use of their personal data for AI training as a condition for providing services.
- The IMDA's guidelines include a standardized info card that consolidates key information about AI chatbots, similar to a nutrition label.
- The guidelines aim to help consumers make more informed decisions when using AI applications and establish a consistent approach to transparency across different chatbot services.
- Several organisations, including Google and Meta, plan to reference the IMDA guidelines over the next six to twelve months to enhance transparency practices.
- The IMDA guidelines complement existing AI governance frameworks from the Monetary Authority of Singapore and the Ministry of Health, which emphasize transparency in AI deployment.
Singapore's new guidelines require organizations to inform consumers when their personal data is used for training generative AI models, effective July 20. The Personal Data Protection Commission (PDPC) emphasizes that consent for data use cannot be a condition for service provision beyond reasonable limits.124
The guidelines state that if anonymized data is used, notifications are not necessary. Organizations cannot deny services to consumers who refuse data use for AI training. Toh Ming Min, a director at BR Law Corporation, noted, “If the individual declines to allow his personal data to be used for AI training, the baseline service may have no real value.”
In conjunction with these regulations, the Infocomm Media Development Authority (IMDA) has rolled out a transparency card for generative AI chatbots, designed to provide users with clear information about the capabilities and limitations of these technologies. This card will help users understand how their data is handled and how to report issues.3
The IMDA's guidelines are among the first globally and aim to establish a consistent framework for transparency in AI applications. They were developed after consultations with various stakeholders, including technology companies and the public. Several organizations, including Google and Meta, plan to adopt these guidelines within the next year.7
These initiatives are part of Singapore's broader efforts to enhance AI governance and set international standards for responsible AI deployment, addressing concerns around data privacy and user safety.
“Under the new rules, organisations must notify consumers if personal data is used for AI training, but are not required to offer opt-out mechanisms. IMDA's voluntary framework, which includes a standardized 'chatbot info card,' has been adopted by Google, Meta, DBS, and Singapore Airlines within the next six to 12 months.”
