AI news ·
IMDA and Microsoft sign MOU to strengthen AI safety and security
IMDA and Microsoft signed an MOU to collaborate on AI safety. The partnership will produce benchmarks, access frameworks, and a white paper on testing frontier models.
The Infocomm Media Development Authority (IMDA) and Microsoft Corporation signed a memorandum of understanding last Friday to deepen their joint work on artificial intelligence safety and security. The agreement, announced in a joint statement, targets three tracks: technical and research collaborations, information sharing on AI safety, and a policy framework to give governments and infrastructure operators trusted access to frontier AI models.
The partnership arrives as AI systems become more capable and autonomous, raising the stakes for evaluation and risk mitigation that no single organisation can handle alone.
Three pillars of the collaboration
The first pillar covers joint research into agentic AI and the development of evaluation methods, tools, and benchmarks. That includes a focus on multilingual AI safety, ensuring models perform safely across different languages. The partners also plan to study how to strengthen societal resilience against the trust and safety challenges posed by AI systems.
For information sharing, IMDA and Microsoft will exchange knowledge, best practices, governance frameworks, and research findings. Where it makes sense, that intelligence may be extended to relevant ecosystem partners.
The final track involves Singapore's AI Safety Institute, other government agencies, and infrastructure operators. Together they will explore how to structure access to frontier AI models for safety and security testing. This work will produce a white paper that examines both the demand-side needs of public agencies and the supply-side policy considerations for model providers.
What the leaders said
Kiren Kumar, Deputy Chief Executive of IMDA, said the partnership shows how government and industry can scale good governance for the public good.
"Our collaboration on AI safety with Microsoft demonstrates how government and industry can partner together to drive and scale good governance for the public good," Kumar said. "This goes beyond developing policy frameworks towards jointly building benchmarks, tools and other evaluation methods, concretely advancing the state of evaluation sciences in an area of increasing importance."
Natasha Crampton, Microsoft's Chief Responsible AI Officer, framed the deal as a way to combine public-sector insight with private-sector technical and operational muscle. She leads Microsoft's responsible AI practices.
"Singapore is helping shape the global conversation on how AI can be developed and deployed safely, securely, and responsibly," Crampton said. "Through this partnership with IMDA, we can combine government insight with Microsoft's technical and operational experience to strengthen AI evaluation, address emerging risks, and build greater trust in advanced AI systems."
Why this matters for IT and development
The MOU isn't just a diplomatic handshake. Its deliverables-new benchmarks, access frameworks, and a white paper on frontier-model access-will directly influence how enterprise IT teams and developers evaluate and deploy advanced AI in regulated environments. As agentic AI moves from research to production, the safety and compliance tools that emerge from this collaboration are likely to shape internal risk assessments, red-teaming workflows, and procurement checklists.
For IT leaders and policy-facing professionals building governance structures inside their organisations, staying ahead of these frameworks is essential. Resources such as the AI Learning Path for Policy Makers can help translate the principles being forged between IMDA and Microsoft into concrete, in-house governance practices.