Singapore's Agentic AI Ambitions Hinge on Code Verification, Says Industry Expert
Source: TNGlobal
Singapore has positioned itself as a global leader in agentic AI governance by releasing one of the world's first dedicated frameworks for AI agents. But the real test, says industry experts, lies in whether the code those agents produce can be trusted — and a layered verification approach combining deterministic analysis with AI review may be the answer.
Singapore has positioned itself as a global leader in agentic AI governance by releasing one of the world's first dedicated frameworks for AI agents that can plan and execute software tasks autonomously. But the real test of that ambition, according to industry experts, is whether the code those agents produce can be trusted.
At May's ATxSummit, Singapore refreshed its Model AI Governance Framework for Agentic AI, first published in January. The framework is blunt on accountability: an AI agent may do the work, but the people who deployed it own the result. Singapore ranks among the world's leading adopters of agentic AI, and a growing share of software shipping across the region is now written in whole or in part by a machine.
The problem, as Amazon CTO Werner Vogels articulated at AWS re:Invent, is a growing "verification debt" — AI produces code faster than teams can verify it. Marcus Low, Corporate Vice President at Sonar and Managing Director for APJ, argues in a TNGlobal Insider piece that the solution lies in a layered verification model combining deterministic analysis (fast, repeatable, auditable) with AI-assisted review (good for logic and context) and automated testing, with human oversight as the final layer.
Sonar developed its approach in partnership with IMDA and using technology acquired from the National University of Singapore. Its remediation agent proposes fixes to code issues and checks them against the same analysis before opening a pull request for developer approval. The firm's research found that teams using its automated code analysis for verification are 44 percent less likely to suffer outages caused by AI-generated code.
Why it matters for Singapore: The city-state's ambitions as a trusted hub for AI development depend on solving the verification challenge. As businesses across the region increase their reliance on agentic coding tools, Singapore's governance framework — backed by partnerships between government agencies like IMDA, research institutions like NUS, and industry players — could become a blueprint for how to balance AI-driven speed with accountability. Getting verification right means this year's productivity gains don't become next year's rework and downtime.
