OpenAI opens Applied AI Lab in Singapore; IMDA updates rules

OpenAI opened its first Applied AI Lab outside the United States in Singapore as part of a S$300 million partnership announced at the ATx Summit. OpenAI announced the lab will create more than 200 Singapore-based technical roles over the next few years and position the city-state as a hub for forward-deployed engineers working on AI deployment in public service, finance and digital infrastructure.

The initiative, called OpenAI for Singapore, will fund a local lab focused on deploying AI systems with organizations and building talent through education and workforce programs. OpenAI plans to support a Singapore chapter of the OpenAI Academy, run Codex for Teachers hackathons, join the National AI Impact Programme, and work with the Ministry of Education and GovTech on training. The partnership also includes accelerator-style workshops for micro-entrepreneurs and small businesses on using AI for operations and customer service.

The lab will host forward-deployed engineers who will collaborate with partner organizations on integration, safety practices and operational rollouts. OpenAI stated the S$300 million commitment is intended to cover investment in the lab, programs and local hiring over the coming years. Singapore’s Ministry of Digital Development and Information will coordinate the partnership.

The Infocomm Media Development Authority issued an updated governance framework for agentic AI at the same summit. The revised framework builds on Singapore’s 2020 Model AI Governance Framework and reflects input from more than 60 organizations. It adds guidance on risks linked to multi-agent systems, third-party agents, automation bias and human accountability.

IMDA published more than ten case studies showing how organizations applied the framework. One case study from Dayos describes an AI ticketing agent that handles internal IT requests using tiered risk levels. Low-risk, reversible actions such as password resets are automated with periodic audits, moderate-risk actions require human approval, and higher-risk actions are excluded.

A Tencent case study covers CodeBuddy, an agentic coding assistant that plans, writes and deploys code through natural-language instructions. CodeBuddy enforces preset defaults and configurable permissions and requires human approval for file edits, shell commands, network requests or external tools. The system explains complex commands in plain language before approval and flags suspicious commands for review.

GovTech Singapore documented a phased rollout of agentic coding assistants within government. The first phase limited use to GovTech employees, disallowed external tools and restricted the assistants to low-risk systems. GovTech also implemented central logging, a process for connecting approved external tools and testing against potential attacks.

“Singapore’s response to AI includes growing new sectors, anchoring global frontier companies, and equipping workers with relevant skills,” Chng Kai Fong, Permanent Secretary for Digital Development and Information, said.

OpenAI for Singapore will focus on projects aligned with Singapore’s AI priorities and on roles that support deployment, safety testing and collaboration with public and private sector partners. The partnership combines investment, workforce development and operational support for organizations adopting AI.