Singapore has put forward a proposed international standard, ISO/IEC 42119-8, that would set standardised testing approaches for generative AI systems, including benchmarking and red teaming methodologies.

The standard is being considered at the 17th ISO/IEC JTC 1/SC 42 plenary meeting, held in Singapore from April 20 to 24, 2026. Singapore's Infocomm Media Development Authority (IMDA) and Enterprise Singapore (EnterpriseSG) are co-hosting the event, the first time the bi-annual plenary has been held in the ASEAN region.

More than 35 national bodies and over 250 AI experts are attending, representing countries including the United States, the United Kingdom, China, Australia and New Zealand.

ISO/IEC 42119-8 is the first international standard of its kind for testing generative AI systems, according to the agency.

The proposed standard would establish a framework for AI testing that improves the reproducibility and comparability of results across organisations and jurisdictions. It builds on IMDA's earlier domestic work, including the AI Verify Toolkit and a starter kit for testing large language model applications for safety and reliability.

International ISO standards typically undergo a multi-year review and ballot process before formal publication.

The standard is part of Singapore's broader push on AI governance, which includes its AI Safety Institute and leadership of the ASEAN Working Group on AI Governance.

IMDA and EnterpriseSG are running capacity-building workshops alongside the plenary, including sessions for ASEAN member states on AI standards and for national standards bodies on applying ISO/IEC AI standards in governance.

https://www.imda.gov.sg