- Singapore has proposed a new international standard, ISO/IEC 42119-8, to standardize testing methods for Generative AI.
- The initiative was launched by IMDA and Enterprise Singapore, discussed at the ISO and IEC meetings from April 20–24, 2026.
- The event gathered more than 35 countries and over 250 AI experts from nations such as the US, UK, Germany, Japan, and China.
- This is the first time this high-level AI standards conference has been held in Southeast Asia.
- The standard focuses on benchmarking and red teaming to evaluate the safety and reliability of Generative AI systems.
- The goal is to create a unified testing method that allows for the comparison and reproduction of results across AI systems.
- This helps increase trust and transparency while promoting the safe deployment of AI at scale.
- IMDA leadership emphasized that standards are the “silent infrastructure” that helps all parties in the AI ecosystem operate effectively together.
- Three main priorities include: keeping pace with AI development, ensuring inclusivity, and enhancing system testing.
📌 Conclusion: Singapore is leading the way in building global standards for Generative AI with ISO/IEC 42119-8, bringing together over 250 experts and 35 countries. By standardizing testing and evaluation, this initiative increases trust and comparability between AI systems. This could become a crucial foundation for promoting safe, transparent, and large-scale AI deployment worldwide.
