Singapore strengthens AI governance with new guidelines

Singapore aims to strengthen its artificial intelligence (AI) governance to create a business-friendly ecosystem. The Ministry of Digital Development and Information (MDDI) is rolling out new guidelines to build trust in AI-enabled products and services, which is essential to sustaining growth and innovation in the sector.

During Data Privacy Week, MDDI announced a number of key initiatives to enhance transparency and security in AI. One of the key measures is the introduction of security labels for generative AI applications, providing users with clear information on the appropriate use, risks and testing of AI models. The guidelines aim to ensure transparency and understanding of the functioning and limitations of generative AI.

In addition, new guidelines under AI Verify establish a baseline for transparency and testing in AI development, addressing issues of misinformation and bias in AI output. The Infocomm Media Development Authority (IMDA) will consult with the industry to create a trusted AI environment.

To support enterprises, MDDI is releasing a guide to generating synthetic data to help create realistic data sets for training AI models without compromising privacy. Synthetic data, an important component of privacy-enhancing technologies, offers a solution to the privacy challenges needed for AI innovation.

The ASEAN Data Anonymization Guide, scheduled for release in early 2025, will help enterprises in the region with data anonymization methods while ensuring secure data transfer across borders. The guide aims to improve the digital ecosystem in the region.

Experts at the panel highlighted the unique data management challenges in generative AI and the importance of diverse data sets and robust data protection throughout the AI ​​lifecycle. They also noted the need to educate consumers on the implications of data sharing to make informed decisions.