By James Konstantin Galvez ( April 20, 2026, 06:33 GMT | Insight) -- Singapore is pushing for faster global AI governance by promoting earlier standard-setting and stronger testing frameworks, as officials warn standards must keep pace with rapid technological change. At an international plenary, authorities highlighted gaps in testing and assurance while proposing ISO/IEC 42119-8, a new standard for benchmarking and red teaming generative AI systems. Officials said credible certification and inclusive participation are key to adoption, underscoring growing alignment between technical standards and regulatory policy to support safe, interoperable cross-border AI deployment.Singapore is calling for a faster, more integrated approach to global AI governance, with officials emphasizing that standards must evolve in tandem with technological development and be backed by credible testing and certification frameworks to remain relevant.Speaking at a plenary meeting* in Singapore on Monday, Infocomm Media Development Authority Chief Executive Ng Cher Pong framed standards as foundational to scaling artificial intelligence safely across borders.“Standards are the quiet infrastructure that enables interoperability, consistency, and trust at scale,” he said, warning that in the fast-moving AI landscape, “the standards-setting process cannot afford to move at a glacial pace.”Ng highlighted testing and assurance as a key gap in the current global framework, pointing to Singapore’s push to formalize methodologies such as benchmarking and red teaming. He said such efforts are “critical standards that will form the foundation of more trustworthy testing and facilitate more reproducible and comparable results.”Enterprise Singapore Director-General Choy Sauw Kook reinforced the need to embed standards earlier in the innovation cycle, arguing that governance frameworks must keep pace with emerging technologies rather than follow them.“Standards can no longer follow mature markets — they must evolve alongside emerging technologies,” she said, adding that early standards act as “guardrails that allow startups and enterprises to innovate with confidence.”Choy also stressed that adoption depends on robust conformance systems, including certification and accreditation. “Developing standards is only the starting point. Impact is measured by adoption. And adoption depends on credible testing, certification and accreditation,” she said.Against this backdrop, Singapore has proposed ISO/IEC 42119-8, a new international standard aimed at establishing common testing methodologies for generative AI systems, with a focus on benchmarking and red teaming (see here).The proposal seeks to improve the reproducibility and comparability of AI testing outcomes, addressing a key barrier to trust and deployment.The initiative builds on Singapore’s domestic efforts, including the AI Verify toolkit and the Global AI Assurance Sandbox, which generate real-world testing insights to inform the development of standards.Officials also underscored the need for more inclusive participation in standards-setting, particularly from emerging economies, to ensure that frameworks reflect diverse use cases and remain globally applicable.At the same time, they highlighted the growing convergence between technical standard-setting and regulatory policy, as governments seek interoperable systems that can support cross-border AI deployment while maintaining safety and trust.*17th ISO/IEC JTC 1/SC 42 Plenary Meeting, Singapore, April 20-24, 2026Please email editors@mlex.com to contact the editorial staff regarding this story, or to submit the names of lawyers and advisers....