OpenAI is set to open its inaugural Applied AI Lab beyond the United States in Singapore. This initiative is part of a new collaboration with the Ministry of Digital Development and Information.
Revealed at the ATx Summit, the project, named OpenAI for Singapore, is supported by funding exceeding S$300 million.
Over the coming years, the facility will generate more than 200 technical roles positioned in Singapore. OpenAI highlighted that Singapore will also serve as a global hub for its on-site engineers who will assist organizations in deploying AI. The lab’s activities will align with Singapore’s AI Mission goals, including public service, finance, and digital infrastructure.
Focus on deployment and talent
The company will collaborate with governmental agencies and local partners on education and workforce development programs through the Ministry of Education and GovTech. OpenAI intends to support teachers via a Singapore section of the OpenAI Academy, participate in the National AI Impact Programme, and host Codex for Teachers hackathons.
The collaboration includes strategies to work with local partners to launch accelerator programs for AI-native startups, offering workshops for micro-entrepreneurs and small business owners on how leaders and SMEs can integrate AI in their operations and customer relations.
Chng Kai Fong, Permanent Secretary for Digital Development and Information, stated that Singapore’s strategy to AI involves expanding new industries, securing global frontier companies, and training workers with appropriate skills.
Singapore updates its agentic AI guidelines
Singapore has refreshed its governance structure for agentic AI, initially introduced by the Infocomm Media Development Authority at the World Economic Forum in January 2026. This framework builds upon Singapore’s existing Model AI Governance Framework for AI from 2020, providing organizations with guidance on the responsible deployment of AI agents, including risk mitigation measures inherent in agentic AI.
IMDA has now refined this framework after gathering input and case studies from the sector, incorporating feedback from over 60 organizations, such as AWS, DBS, Google, and Salesforce.
The updated guidance addresses risks related to multi-agent systems, third-party agents, automation bias, and human accountability. More than ten case studies are now included to illustrate how organizations have implemented these guidelines.
Contributions for case studies came from local Singapore-based and global organizations, including Ant International, City Developments Limited, Cyber Sierra, Dayos, Google, Knovel, OCBC, PwC, Stability Solutions, Tencent, Terminal 3, Workday, X0PA, and GovTech Singapore.
Case studies showcase governance controls
One case study features Dayos, an enterprise AI automation firm headquartered in Singapore operating in the US. Dayos created an AI ticketing agent managing internal IT requests. The agent automatically resolves certain requests and forwards others to human agents when necessary.
Dayos utilized tiered risk levels to define permissible actions for the agent. Low-risk, reversible tasks, such as password resets, were automated with biweekly audits. Moderate-risk actions required human approval before execution. High-risk tasks, like permission alterations that are difficult to reverse, were excluded from the agent’s authority.
Tencent’s case study examines CodeBuddy, an agentic AI coding system developed by Tencent Cloud. CodeBuddy can plan, write, and deploy code using natural language instructions and can access file systems, terminal commands, external APIs, and MCP tools.
CodeBuddy operates on preset defaults and adjustable permissions. Human approval is mandatory for activities like file editing, shell command execution, network requests, or external tool usage.
Before user approval, the system explains complex commands in straightforward language. Suspicious commands still require manual approval, even if similar commands were previously authorized.
GovTech Singapore’s case study outlines the introduction of agentic coding assistants in the government. The initial phase was restricted to GovTech employees, limited to internal tools only, and focused on low-risk systems. GovTech established centralized logging and a framework for incorporating approved external tools. The agency also screened the system for potential vulnerabilities.
(Photo by Mike Enerio)
See also: GPT-5.5 is OpenAI’s most capable agentic AI model yet
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The extensive event is part of TechEx and co-hosted with other premier technology events; click here for further information.
AI News is powered by TechForge Media. Explore various upcoming enterprise technology events and webinars here.
