On February 27, the seminar themed “Technology Diffusion and Asia Prosperity” was held in Singapore. The event was hosted by Singapore-based consultancy Edge Research and marked the launch of its “Asia Prosperity Initiative.” Experts and representatives from policymaking, academia, and industry gathered to engage in in-depth discussions on trends in AI technology diffusion, the development of AI agents, and the future of regional AI governance.
Technology Diffusion Drives Social Transformation; Youth Development Pathways Being Rewritten
During the roundtable session on “How Responsible Technology Diffusion can enable a Prosperous Asia”, participants focused on the impact of AI development on employment structures, education systems, and talent cultivation models.
Professor Lawrence LOH, Director of Centre for Governance and Sustainability at the National University of Singapore Business School, noted that the main arena of frontier technological innovation has gradually shifted from traditional universities to corporations. Technology companies such as Google in the United States, Alibaba and Tencent in China have become key engines of AI innovation. Drawing on his teaching observations, he pointed out that certain job positions are facing displacement pressure from AI, intensifying competition for university graduates. How education systems can cultivate new types of talent suited to the AI era has become an urgent and practical question.
ZHANG Fan, Founder of Edge Research, cited U.S. studies indicating that generative AI has a greater impact on entry-level positions than on mid- to senior-level roles. In this context, companies will play an increasingly critical role in talent training and workplace adaptability. He referenced the example of U.S. tech company Palantir directly recruiting high school graduates for systematic training, describing it as an innovative exploration beyond traditional academic pathways and a new approach to diversified talent development.
Rapid Deployment of AI Agents; Heightened Security and Accountability Challenges
During the roundtable session on “AI for Good and Development in Southeast Asia”, discussions shifted toward the development and governance of AI agents.
Alvin CHIA, Head of APAC Digital Asset Innovation at Northern Trust, observed from the financial sector perspective that AI agents demonstrate significant efficiency advantages in areas such as customer service automation, risk management assistance, and trading decision support. However, the high sensitivity of financial systems necessitates extreme caution in deploying AI agents. “Each new tool integrated into an AI agent adds a potential attack surface,” he noted. Unauthorized agent operations may trigger compliance risks and financial fraud issues, making standardized protocols and cybersecurity mechanisms particularly critical.
HUANG Jingyang, Assistant Professor at the School of Public Policy, The Chinese University of Hong Kong (Shenzhen), highlighted that some AI hardware and agents currently achieve cross-application operations through underlying permissions such as screen reading and simulated clicks. Originally designed to assist visually impaired users, such technologies in commercial applications may lead to excessive privacy concessions and data misuse risks, while also disrupting existing internet business models. In contrast, cross-application collaboration mechanisms based on protocols such as A2A (Agent-to-Agent) and MCP, though still requiring user experience optimization in the short term, offer greater robustness in terms of privacy protection and ecosystem sustainability.
WANG Yin, Assistant Professor of Accounting at Singapore Management University, emphasized that the “black-box” decision-making nature of AI agents complicates accountability. When system decisions go wrong, it remains unclear whether responsibility should lie with developers, deployers, end users or other possible stakeholders. No country has yet provided a definitive answer. Therefore, defining authorization boundaries and establishing traceable accountability mechanisms will be central issues in designing AI agent governance frameworks.
ZHANG Fan added that AI agents are akin to “butlers”. If individuals hand over even the keys to their bedrooms or safes to AI agents, personal privacy and security are inevitably at risk. From this perspective, AI agent applications must strike a balance among convenience, reliability, and security.
Ongoing Governance Exploration; Promising Prospects for Southeast Asia
Participants also exchanged views on global AI governance trends and Southeast Asia’s regional development trajectory.
Benjamin GOH, Senior Assistant Director at Singapore’s National AI Group, remarked that the European Union’s AI Act, as an early mover in comprehensive AI regulation, has exerted significant influence on global AI governance debates. Meanwhile, Asian countries are actively exploring governance pathways suited to their respective development stages. The recently held India AI Impact Summit put forward proposals on issues such as data sharing and management. This January, Singapore’s Infocomm Media Development Authority (IMDA) released the Model AI Governance Framework for Agentic AI, aiming to provide a reference framework for industry and promote international dialogue.
WANG Yin stressed that at the governmental level, differences in institutional capacity across Southeast Asian countries must be acknowledged. Beyond regulation, effective implementation and policy execution deserve equal attention. HUANG Jingyang added that from an industry perspective, self-regulation and consensus on standards are particularly important. Before hard regulations are introduced, industries are often the first to identify emerging risks; therefore, establishing voluntary standards within the sector is essential to advancing effective governance.
Regarding AI development prospects in Asia, particularly Southeast Asia, participants expressed optimism. Although Southeast Asian countries may follow in foundational large-model capabilities, they possess distinct strengths in application-layer innovation. With advantages in population scale, market diversity, and geostrategic positioning, Southeast Asia is well-positioned to build upon high-quality global foundation models, advance deep localization and adaptation, accelerate inclusive AI applications across sectors, and drive the regional digital economy toward a more inclusive growth trajectory.