Intelligent agents and agentic AI for ASEAN transformation and Asia’s future
· 22 min read
This article is the third part of a four-piece series on the rise of AI agents. You can find part one here and part two here.
Asia and ASEAN must take a proactive and calculated approach in order to fully realise the transformative potential of agentic AI for economic growth and sustainable development:
Asia and ASEAN must strategically invest in research and development (R&D) if they want to become significant participants in the agentic AI arena. Governments, educational institutions, and the private sector must collaborate to support local innovation, nurture talent, and build a robust ecosystem for AI development. The importance and potential outcomes of such initiatives are demonstrated by compelling examples and verified facts.
Asia's governments are forming national agendas and allocating substantial financial resources to research and development as they realise the strategic importance of AI.
• China: As a worldwide leader, China has invested heavily in AI through initiatives like the "Next Generation Artificial Intelligence Development Plan." With the aim of positioning China as a global leader in artificial intelligence by 2030, this strategy includes significant funding for basic research, vital technologies, and talent development. According to the Chinese Ministry of Science and Technology, $37.8 billion was spent on basic research in 2023, with a significant portion of the money going towards artificial intelligence and related fields.
• South Korea: A goal to become a global AI powerhouse is outlined in South Korea's "AI Strategy 2030" with specific investments in R&D, talent development, and infrastructure development. To support AI research in several areas, the government has set aside billions of dollars. The Ministry of Science and ICT, for instance, pledged a $1.6 billion investment in AI research and development over the next five years (2024–2028), with a focus on basic AI technologies and their applications.
• Singapore: As stated earlier, Singapore's Smart Nation strategy heavily relies on AI research and development. The AI Singapore (AISG) initiative, which promotes collaborations between government, industry, and academia to develop and apply AI solutions, is one of the major initiatives funded by the National Research Foundation (NRF) Singapore. More than 100 AI projects in a range of industries have been financed by AISG in an effort to foster innovation and build regional AI capabilities.
• Japan: Japan's "AI Strategic Council" has outlined a plan to enhance AI R&D capabilities, with a focus on areas including medical AI, manufacturing AI, and natural language processing. The government has increased its R&D spending for AI-related projects in an attempt to regain its competitive edge in the global AI industry. For example, the New Energy and Industrial Technology Development Organisation (NEDO) actively encourages cutting-edge AI research projects.
Table 1: Summary of the various regional AI initiatives
World-class universities and dedicated research institutions play a crucial role in fundamental AI research and talent development.
• Tsinghua University (China): The Institute for Artificial Intelligence at Tsinghua University, which is regularly ranked as one of the top universities in the world for computer science and AI, is at the forefront of cutting-edge research in areas including machine learning, computer vision, and natural language processing. China's AI capabilities are significantly improved by the calibre of its graduates and the volume of research they generate.
• KAIST (South Korea): Strong robotics and artificial intelligence research programs are hallmarks of the Korea Advanced Institute of Science and Technology (KAIST). Its AI Research Centre is working on several cutting-edge initiatives that foster innovation and produce highly skilled AI experts.
• National University of Singapore (NUS) and Nanyang Technological University (NTU) (Singapore): Strong computer science degrees with a focus on artificial intelligence research are offered by these two Singaporean universities. They actively collaborate with business and government to promote basic and applied AI research through initiatives like AI Singapore. For instance, NUS's AI Lab works on subjects like explainable AI and dependable machine learning.
• The University of Tokyo (Japan): Japan advances in areas like robotics and natural language processing thanks to the University of Tokyo's Graduate School of Information Science and Technology, which conducts cutting-edge research in a range of AI domains.
These prestigious universities' existence and robust research initiatives offer a crucial basis for the advancement of agentic AI in Asia.
The private sector is a significant driver of AI R&D, translating fundamental research into practical applications and investing heavily in developing proprietary AI technologies.
• Technology Giants: Companies like Alibaba, Tencent, Baidu (China), Samsung, LG (South Korea), and SoftBank (Japan) have set up dedicated AI research laboratories and are spending billions of dollars to develop advanced AI capabilities, especially those pertaining to agentic AI. For example, Alibaba's DAMO Academy conducts cutting-edge research in domains including computer vision and natural language processing, with applications ranging from e-commerce to cloud computing.
• AI Startups: A flourishing ecosystem of AI companies is emerging in Asia, focussing on specialist applications and advancing innovation in specific AI domains. By regularly collaborating with bigger organisations and corporations, these companies add to the broader field of AI research and development. For instance, AI companies have sprung up all around Singapore, attracting venture capital investment and fostering innovation in industries like fintech, healthcare, and smart cities. According to a Dealroom.co analysis, Southeast Asian AI companies raised over $1 billion in 2024, indicating the private sector's growing interest and investment.
• Industry-Specific R&D: Numerous industries are spending money on AI research & development in order to enhance their operations and produce new products. In the automotive industry, for example, companies such as Toyota and Hyundai are investing heavily in the development of autonomous driving technology, which is a critical application for advanced AI agents. AI is also being used by pharmaceutical companies to develop and discover novel medications in the healthcare sector.
The business sector's active participation and substantial investment in turning AI research findings into useful applications are essential for the adoption of agentic AI across ASEAN and Asia's industries.
Promoting collaboration between academia, industry, and government is essential to accelerating AI research and development and ensuring that studies are in step with the challenges and needs of the real world.
• Joint Research Projects: Governments can encourage corporations and universities to collaborate on research projects by providing financial incentives and cooperative platforms. Singapore's AI Singapore initiative serves as an illustration of this, bringing together academics from universities and business partners to address AI issues on a national scale.
• Industry Consortia: By promoting the sharing of resources, information, and expertise, industrial consortia focused on specific AI applications can hasten innovation and adoption. For example, consortia can focus on developing and deploying automation solutions driven by AI in the manufacturing sector.
• International Collaboration: International partnerships can facilitate access to global resources and expertise, encouraging the exchange of ideas and accelerating the development of AI. ASEAN members can explore joint research projects with leading AI nations and participate in global AI research initiatives.
The establishment of a regional AI research network within ASEAN that connects leading universities and research institutions across member states could foster collaboration on significant Agentic AI research topics relevant to the region's particular challenges, such as climate change adaptation and sustainable agriculture. This network may promote data and resource sharing, researcher exchanges, and collaboration on research efforts.
According to UNESCO's Science Report 2021, East and Southeast Asia's share of global R&D investment has been steadily increasing, reaching over 40% in 2018. This demonstrates the region's increasing commitment to the advancement of science and technology, especially AI.
ASEAN and Asia must make strategic investments in R&D, foster collaboration, and develop talent if they are to keep up with the global AI revolution and become leaders in the development and use of agentic AI for a sustainable and prosperous future. This requires targeted funding, steady commitment, and a collaborative mindset from all stakeholders.
Strong ethical and legal frameworks must be established in order to navigate the complex societal and economic effects of agentic AI. These structures should encourage innovation while lowering risks and ensuring fairness, transparency, accountability, and safety in the development and use of these powerful technologies. New regulatory initiatives and verified statistics underscore the growing worldwide awareness and efforts in this vital area.
Ethical principles, which deal with concerns about bias, privacy, autonomy, and misuse potential, offer a vital foundation for the responsible development and use of AI.
• Addressing Algorithmic Bias: AI systems, notably agentic AI, have the potential to inadvertently reproduce and magnify biases in the data they are trained on, leading to unfair or discriminatory outcomes. Research has shown that bias in facial recognition systems disproportionately affects certain demographic groups. For example, a 2018 study by MIT Media Lab discovered that commercial facial recognition algorithms had significantly higher error rates for women with darker skin tones than for white men. Ethical guidelines should highlight the need for representative and diverse datasets, comprehensive techniques for identifying and reducing bias, and ongoing monitoring of AI system outputs.
• Ensuring Data Privacy and Security: Agentic AI systems commonly use large amounts of data, which presents significant privacy and security challenges. Laws like the General Data Protection Regulation (GDPR) of the European Union have created data protection principles, including the right to privacy, data minimisation, and purpose limitation. Similar frameworks are being developed or are being considered by other Asian countries. The Personal Data Protection Act (PDPA) of Singapore, for instance, lays forth requirements for how businesses must collect, use, and disclose personal data. The ethical criteria for agentic AI should be based on these legal frameworks and should give priority to data anonymisation, secure data management practices, and transparency regarding data use.
• Managing Autonomy and Control: Control and accountability issues arise when agentic AI systems become more autonomous. Clear ethical guidelines are needed to establish the appropriate degrees of autonomy for different applications, assign responsibility for the actions of AI agents, and ensure human oversight in critical decision-making processes. The "human-in-the-loop" theory, which preserves human operators' capacity to step in and reverse AI decisions, is at the centre of these issues.
• Preventing Misuse: Ethical frameworks also need to address the potential for dangerous applications of agentic AI technologies, including the development of autonomous weapons or the use of AI in cybercrime. It may be necessary to establish international agreements and collaboration in order to establish standards and prohibit such harmful applications.
Legislative frameworks pertaining to AI at various levels, including agentic AI, are being developed by governments worldwide. The landscape is dynamic and always shifting, with different areas employing different tactics.
• European Union's AI Act: The EU's AI Act, which is expected to be fully implemented in the coming years, uses a risk-based approach to categorise AI systems based on their potential harm and impose regulations appropriately. Strict guidelines for data governance, transparency, precision, and human supervision will apply to high-risk AI systems, which might include some Agentic AI applications in critical sectors. This law establishes the benchmark for comprehensive AI laws.
• United States: The US has adopted a more sector-specific approach to AI regulation, with a number of institutions issuing recommendations and rules relevant to AI applications within their own areas (such as healthcare and finance). The White House has released an AI Bill of Rights that provides rules for the proper creation, usage, and implementation of automated technologies.
• China's Regulations: China has created legislation focused on AI, such as those governing algorithmic suggestion management and the ethical oversight of next-generation AI. These regulations reveal China's focus on finding a balance between innovation, control, and societal stability.
• ASEAN Member States: Several ASEAN member states are actively exploring and developing AI regulatory frameworks.
• Singapore: Initially, Singapore has taken a pragmatic and creative approach, prioritising the creation of guidelines and principles over strict rules. To promote the appropriate use of AI, the Infocomm Media Development Authority (IMDA) has published a Model AI Governance Framework and a Compendium of Use Cases.
• Malaysia: Legal and ethical frameworks for AI are taken into consideration in Malaysia's National AI Roadmap.
• Thailand: Thailand's National AI Strategy also highlights the importance of ethical concerns and the development of appropriate regulations.
• Indonesia: It is expected that Indonesia's early stages of creating a national AI strategy will include discussions on ethical and regulatory issues.
According to a 2023 Brookings Institution study, more than 60 countries have released national AI plans, many of which include sections on ethics and governance. This implies that there is widespread recognition of the need for responsible AI development.
Effective regulatory frameworks for Agentic AI in ASEAN and Asia should consider the following key elements:
• Risk-Based Approach: A risk-based approach that separates regulatory requirements depending on the potential impacts of different agentic AI applications, such to the EU's AI Act, can be effective. High-risk applications in industries including law enforcement, healthcare, and finance would be subject to stricter regulations than low-risk ones.
• Transparency and Explainability: Regulations that promote transparency in the design and operation of agentic AI systems should support the development of Explainable AI (XAI) techniques that help humans understand how AI agents make decisions. This is crucial for ensuring responsibility and building trust.
• Accountability and Responsibility: Different lines of accountability and responsibility must be established for the acts of agentic AI systems. This includes identifying who is responsible for ensuring that these agents act in a safe and ethical manner: creators, deployers, or users.
• Human Oversight and Control: In accordance with legislative frameworks, key applications of agentic AI, particularly those that could significantly impact society, should be supervised and controlled by humans. This could mean requiring person-in-the-loop procedures or establishing clear protocols for human intervention.
• Data Governance and Privacy: Strong data governance frameworks that follow international best practices, including the GDPR (or PDPA for Singapore’s context) requirements, are necessary to protect personal data used by agentic AI systems. This includes the requirements for data permission, data security, and data minimisation.
• Promoting Innovation: It is important to develop regulatory frameworks that promote innovation while lowering risks. Excessively prescriptive regulations could impede the development and use of beneficial agentic AI technologies. It is crucial to employ a comprehensive approach that encourages experimentation while staying within ethical bounds.
• Adaptability and Future-Proofing: Given how rapidly AI is advancing, regulatory frameworks need to be adaptable and future-proof to handle changes as the technology develops. This might mean principles-based laws that provide a general framework rather than incredibly specific rules.
• Regional Harmonisation: Standardising AI regulatory principles and norms at the ASEAN level can foster cross-border innovation and trade while ensuring a consistent approach to ethical issues across the region.
Singapore's approach to developing a Model AI Governance Framework, which provides organisations deploying AI solutions with guiding principles and workable implementation processes, can serve as an example for other ASEAN member states. Human supervision, explainability, transparency, and justice are all highly valued in this method.
ASEAN might establish a regional working group on AI ethics and governance. This body may be tasked with developing a framework for regulating AI and a set of universal ethical principles that member states might adapt to suit their particular national contexts. This collaborative approach would promote consistency and facilitate the appropriate development and use of agentic AI in the area.
Establishing cautious and adaptable ethical and legal frameworks is more critical for fostering public trust and fostering an environment where Agentic AI may be employed for the good of Asia and ASEAN overall than it is for lowering dangers. By proactively addressing ethical challenges and establishing clear norms, the region can pave the way for a responsible and substantial AI-driven future.
Strategic public-private partnerships (PPPs) are essential to maximising the benefits and accelerating the adoption of agentic AI across Asia and ASEAN. By combining the resources, expertise, and adaptability of the public and private sectors, these collaborations can remove barriers to innovation, advance the development of practical AI solutions, and ensure their widespread acceptance for the good of society and the economy. Verified facts and successful examples show how revolutionary these partnerships can be.
PPPs in the realm of Agentic AI offer several key advantages:
• Resource Mobilization: While governments can provide infrastructure, funding, and access to public data, the private sector can provide technological know-how, R&D capabilities, and market insights. AI solutions can be developed and put into use more quickly by pooling resources than either sector could by itself.
• Risk Sharing: PPPs enable the implementation of large-scale, potentially transformative initiatives by allowing the operational and financial risks of AI development and implementation to be shared.
• Expertise Integration: Cooperation among academic institutions, government researchers, and AI specialists in the commercial sector fosters the sharing of ideas and the development of more dependable and practical AI solutions.
• Market Access and Scalability: Private sector partners often have proven commercial channels and the ability to scale AI solutions effectively, ensuring that the benefits of publicly funded research are realised by a wider audience.
• Alignment of Objectives: Properly planned PPPs can align the private sector's ambitions for innovation and commercialisation with the public sector's objectives, which include sustainability, economic growth, and societal benefit.
Several successful AI PPP initiatives around the world demonstrate the effectiveness of this collaborative approach:
• AI Singapore (AISG): It was previously noted that AISG is a great example of a successful PPP. It unites the National Research Foundation (NRF), government organisations, industrial firms, and universities in Singapore to encourage AI innovation and implementation in key sectors. AISG funds and supports collaborative AI initiatives, develops a robust AI ecosystem, and strengthens national AI capabilities. The initiatives include AI-powered automotive, healthcare, and urban mobility solutions. For instance, AISG has worked with local hospitals to develop AI tools for more accurate and efficient disease diagnosis.
• The UK's AI Sector Deal: This government-led initiative, which has a significant corporate component, involves significant expenditures in AI R&D, infrastructure, and talent development. In order to position the UK as a worldwide leader in AI, it entails partnerships to build AI innovation networks and support AI adoption across industries.
• Canada's CIFAR (Canadian Institute for Advanced Research) Pan-Canadian AI Strategy: As part of this national policy, the federal government, leading academic institutions, and the private sector collaborate to advance AI research, cultivate talent, and promote the responsible application of AI. CIFAR grants AI research chairs across the country and promotes collaboration between AI researchers and industry partners.
• EU's Digital Innovation Hubs: PPPs are often a part of the Digital Innovation Hubs (DIHs) network established by the European Union to help SMEs embrace AI technology and undergo digital transformation. These hubs provide access to technology, information, and finance for businesses looking to integrate AI into their operations.
While governments can act as financiers and conveners, the private sector provides the technological know-how and commercial application capabilities to support AI innovation and adopt
Given the diverse needs and opportunities across ASEAN and Asia, several key areas could benefit significantly from well-designed PPPs focused on Agentic AI:
• Smart and Sustainable Cities: PPPs can assist in the development and implementation of agentic AI solutions for urban issues such as waste management, energy optimisation, traffic control, and public safety. Governments can provide access to urban data and infrastructure, while private companies can provide AI-powered platforms and analytics. The primary objective of a PPP in Jakarta may be to create an AI-powered integrated traffic management system that optimises traffic flow in real-time, reducing pollution and congestion.
• Precision Agriculture and Food Security: Agentic AI solutions for supply chain optimisation, pest and disease detection, and precision farming can be developed through collaboration between agritech companies, governmental agencies, and agricultural research institutes, enhancing food security and sustainability. Commercial sector partners can provide analytics platforms, drones, and AI-powered sensors, while governments can provide pilot project locations and agricultural data. A PPP in Thailand might focus on developing AI-driven drone-based agricultural monitoring systems that maximise fertiliser and water use while providing farmers with the most recent crop health information.
• Healthcare and Wellness: For applications like drug discovery, individualised treatment regimens, remote diagnostics, and elder care, PPPs can speed up the development and application of agentic AI. While telemedicine platforms, AI-powered diagnostic tools, and robotic assistants can be provided by private sector partners, governments can provide regulatory rules and healthcare data. To enhance access to quality healthcare in rural regions, a PPP in Malaysia may focus on developing AI-powered remote patient monitoring devices that enable healthcare providers to monitor patients' health.
• Sustainable Resource Management: The development of agentic AI solutions for pollution control, environmental monitoring, and renewable energy optimisation can be accelerated by collaboration among technology companies, research institutions, and environmental agencies. While private sector partners can provide grid management systems, data analytics platforms, and artificial intelligence-powered sensors, governments can provide environmental data and policy frameworks. A PPP in Vietnam might focus on developing AI-powered solutions for real-time industrial emissions and water quality monitoring, enabling more effective environmental protection.
• Education and Skills Development: AI-powered customised learning platforms and intelligent tutoring systems can be developed using PPPs to help reduce the skills gap and improve educational outcomes. Edtech companies can provide AI-powered learning resources and platforms, while governments can provide curriculum frameworks and educational data. A PPP in Indonesia might focus on developing AI-driven customised learning platforms that adapt to the needs of every single student, improving educational outcomes in remote and underprivileged communities.
To ensure the success of Agentic AI PPPs in ASEAN and Asia, the following considerations are crucial:
• Clear Objectives and Shared Vision: All partners must have a clear understanding of the project objectives and a shared vision for the desired outcomes.
• Effective Communication and Collaboration Mechanisms: Robust communication channels and collaboration mechanisms are essential for the smooth execution of projects.
• Clear Legal and Regulatory Frameworks: A supportive legal and regulatory environment is necessary to facilitate PPPs and address issues such as data ownership and intellectual property.
• Well-Defined Roles and Responsibilities: The roles and responsibilities of each partner should be clearly defined in the partnership agreement.
• Appropriate Risk and Reward Sharing: The risks and rewards associated with the project should be fairly distributed among the partners.
• Transparency and Accountability: The operations and outcomes of PPPs should be transparent and accountable to the public.
• Long-Term Commitment and Sustainability: PPPs should be designed with a long-term perspective, ensuring the sustainability of the developed solutions.
In order to facilitate PPP for AI, ASEAN might create a regional platform. Bringing together organisations from the public and private sectors, research facilities, and companies interested in collaborating on AI projects would be the main function of this platform. The region may provide resources, best practices, and regulatory frameworks to support the development and successful implementation of AI PPPs.
According to a 2021 World Bank assessment on PPPs in digital infrastructure, well-structured PPPs can lead to a more effective and efficient deployment of technical solutions, drawing on private sector resources and creativity while aligning with public sector goals.
Technology businesses, academic institutions, and environmental authorities can work together to speed the development of agentic AI solutions for pollution control, environmental monitoring, and renewable energy optimisation. Governments can supply environmental data and regulatory frameworks, while private sector partners can supply data analytics platforms, artificial intelligence-powered sensors, and grid management systems. To enable more efficient environmental protection, a PPP in Vietnam might concentrate on creating AI-powered solutions for real-time industrial emissions and water quality monitoring.
Now that the foundational strategies of robust investment, ethical frameworks, and cooperative public-private partnerships have been established, the next critical step for ASEAN and Asia in embracing the era of Agentic AI is building the required human capital and fostering a spirit of regional cooperation. The next and last half of this article will address these crucial enabling factors as well as the need to carefully direct this powerful technology towards the urgent and shared goal of building a sustainable future for the entire region.
illuminem Voices is a democratic space presenting the thoughts and opinions of leading Sustainability & Energy writers, their opinions do not necessarily represent those of illuminem.
illuminem briefings
AI · Public Governance
Aaron Bruckbauer
Pollution · Greenwashing
Jonathan Lishawa
AI · Energy Transition
The Wall Street Journal
Li-ion Battery · Battery Tech
The Economist
AI · Sustainable Lifestyle
South China Morning Post
Public Governance · Green Tech