Singapore is proposing an international standard, ISO/IEC 42119-8, to benchmark and red-team generative AI, aiming for consistent testing, safety, and reliability across markets.

The proposal builds on Singapore’s previous AI assurance work, including the AI Verify Toolkit and testing starter kits for large language models, feeding into broader international standardisation efforts.

The ISO/IEC JTC 1/SC 42 plenary in Singapore will be the first ASEAN regional gathering of this scale, drawing over 250 experts from more than 35 national bodies, including major players like the US, UK, China, Japan, Germany, France, and South Korea.

Singapore has also signaled plans to establish a National AI Council, chaired by Prime Minister Lawrence Wong, to drive large‑scale AI deployment across manufacturing, finance, and healthcare.

The 42119-8 standard discussion is scheduled at the ISO/IEC JTC 1/SC 42 plenary in Singapore from April 20–24, 2026, co-organised by the Infocomm Media Development Authority and Enterprise Singapore.

Beyond the standard itself, the plenary will host capacity-building activities, including AI standards training with ANSI, policy-focused sessions for national bodies, and industry discussions through the AI Assurance Exchange led by the AI Verify Foundation.

The move signals a shift in AI governance toward accountability and measurement, outlining how AI should be tested and governed rather than focusing solely on innovation.

This initiative aligns with Singapore’s goal of becoming a regional hub for trusted AI development, backed by programs like the AI Safety Institute and ASEAN governance frameworks.

ISO/IEC 42119-8 aims to improve reproducibility and comparability of AI evaluation results, strengthening trust and enabling safer deployment of generative AI.