Singapore – Singapore has introduced a proposal for an international standard intended to harmonise the evaluation of generative AI systems, with a focus on benchmarking and red teaming practices.
The draft standard, ISO/IEC 42119-8, seeks to establish consistent testing methods to support the development of reliable AI systems and is set to be discussed at the 17th ISO/IEC JTC 1/SC 42 plenary meeting taking place in Singapore from 20 to 24 April, organised by the Infocomm Media Development Authority and Enterprise Singapore.
The proposed ISO/IEC 42119-8 framework addresses the need for internationally recognised standards as AI technologies continue to expand across industries. It outlines structured approaches to testing generative AI models, aiming to improve consistency in evaluation outcomes and enable meaningful comparison across systems.
By promoting repeatable and transparent assessment processes, the initiative is intended to reinforce confidence in AI technologies and support their safe and responsible deployment.
“We can do more in strengthening the linkages between standards and testing, particularly by translating testing into real-world practice and implementation,” Cher Pong Ng, Chief Executive Officer at Infocomm Media Development Authority, stated.
“Ultimately, standards are not an end in themselves. Their value lies in how they are put to action, in real-world applications and use cases, to solve problems and enhance trust.”
The development of this proposed standard builds on Singapore’s earlier efforts in AI evaluation, including the AI Verify Toolkit and a starter framework for testing LLM-based applications for safety and reliability.
It also aligns with broader initiatives such as the Global AI Assurance Sandbox, which explores emerging practices in AI validation. These efforts complement ongoing participation in international standardisation activities, including the adoption and accreditation of ISO/IEC 42001 and contributions to ISO/IEC TR 24030, which documents practical AI use cases. Collectively, these initiatives contribute to establishing a structured foundation for trustworthy AI implementation.
Alongside the plenary meeting, a series of supporting programmes is being organised to strengthen expertise in AI standards and governance. These include a regional training workshop developed in collaboration with the American National Standards Institute to enhance technical capabilities among ASEAN member states, as well as an international workshop focused on the practical application of ISO and IEC AI standards for policymakers and national standards bodies.
“Embedding standards early is like building new ‘runways’. AI innovation is like a high-performance aircraft; without a safe, well-built runway, its potential remains grounded. Early standards provide the guardrails that allow startups and enterprises to innovate with confidence,” Sauw Kook Choy, Director General of Enterprise Singapore, commented.
“As the AI standards portfolio expands, our conformance infrastructure must not just keep pace; it must lead and remain accessible to all economies regardless of their AI maturity.
Singapore’s broader strategy centres on fostering a dependable AI ecosystem through initiatives such as its AI Safety Institute and active participation in regional governance efforts, including leadership roles within ASEAN working groups on AI governance and digital trade standards.
These collaborations are intended to support the development and deployment of AI systems that meet recognised standards of reliability and accountability for both organisations and end users.
