Singapore Pioneers First International Benchmark for Generative AI Testing with ISO/IEC 42119-8

Singapore has proposed ISO/IEC 42119-8, the world's first international standard specifically for testing generative AI systems.

This new benchmark, focusing on standardized benchmarking and red teaming methodologies, aims to ensure reproducible and comparable results in AI evaluations.

Singapore's Leadership in Global AI Governance

The standard will be discussed at the 17th ISO/IEC JTC 1/SC 42 plenary meeting, hosted in Singapore from April 20-24, 2026, for the first time in the ASEAN region.

Co-organized by the Infocomm Media Development Authority (IMDA) and Enterprise Singapore (EnterpriseSG), the event will convene over 250 experts from more than 35 national bodies worldwide.

This initiative builds on Singapore's history of AI innovation, including the development of the AI Verify Toolkit and the Starter Kit for testing large language model applications.

Previous efforts also encompass the Global AI Assurance Sandbox and national adoption of ISO/IEC 42001 for AI management systems.

Impact on Trustworthy AI Adoption

By establishing a unified framework, ISO/IEC 42119-8 will enhance trust in AI systems, enabling safer deployment across industries and ecosystems.

The standard addresses the rapid proliferation of generative AI, promoting reliability and safety amid growing global usage.

Side events include workshops with the American National Standards Institute to build ASEAN capacity in AI standards.

Future implications point toward broader international collaboration, solidifying Singapore's role in shaping trustworthy AI governance through its AI Safety Institute.

Ultimately, this benchmark lays the groundwork for a more secure and standardized era of generative AI worldwide.