ASEAN’s AI opportunity depends on trustworthy governance

Artificial intelligence (AI) governance cannot succeed if its application is limited to the world’s largest technology markets. 

The rapid advancement of AI raises questions about how societies and governments can capture its benefits, while managing risks that often extend beyond national borders.

AI might create enormous opportunities for ASEAN economies, but it also requires a new approach to regional governance.

Smaller economies, local governments, schools, journalists, and community organisations also require accessible tools, training, and institutional support. 

Concerns around AI governance intensified after a former Anthropic researcher warned on September 8 that AI could pose an existential threat to humanity - concerns that sparked debates during the 81st United Nations General Assembly (UNGA) this September.

AI governance landscape in Asia

China’s AI Safety Governance Framework 3.0, released in September, provides a useful reference for ASEAN discussions.

It proposes a framework based on risk classification, technical safeguards, comprehensive governance, regulatory sandboxes, and international cooperation.

The timing is significant for ASEAN. Southeast Asia is home to diverse economies, languages, legal systems, digital infrastructures, and levels of technological readiness. 

A successful regional approach must support innovation while recognising that AI risks do not stop at national borders. 

Take Malaysia, Indonesia, and Singapore as part of ASEAN countries that currently moving to strengthen its AI governance framework. 

Malaysia through plans for an AI Governance Bill and the establishment of a Malaysia AI Safety Institute. 

Indonesia is preparing two Presidential Regulations: one on a National AI Roadmap for 2026 to 2029, and one on AI Ethics. 

Both finished their inter-ministry harmonisation in 2026 and await the President's signature. 

Singapore, through its 2026 Consensus on Global AI Safety Research Priorities, underscores the need for AI researchers and practitioners to pay close attention to risks such as AI-enabled cyber-attacks.

ASEAN AI communities concerning with AI also initiate aisean.org, a open source web that tracking AI ecosystem progress across ASEAN countries.

Shared opportunities, shared risks

AI can help ASEAN improve productivity, strengthen public services, support small and medium-sized enterprises (SMEs), expand education access, and improve disaster response.

It can assist with multilingual communication, preserve indigenous and traditional knowledge and languages, and help bridge cultures across the region.

Yet, the same systems can create shared risks. 

AI-generated misinformation can spread across borders within minutes, even seconds.

Cyberattacks can become more automated. Personal data can be processed without adequate consent. 

Bias in models disadvantage communities whose languages and cultural contexts are underrepresented in training data, especially for Southeast Asia, which is home to more than 1,561 languages.

The Chinese framework groups AI risks into inherent, application-related, and derivative risks. Inherent risks include hallucinations, data poisoning, prompt injection, bias, poor explainability, and insecure infrastructure. 

Application risks include unauthorised access by AI agents, unsafe tool use, deepfakes, cyberattacks, privacy violations, and disruption to critical infrastructure. 

Derivative risks include employment disruption, environmental pressure, cultural homogenisation, and the widening intelligence divide. 

These categories are relevant to ASEAN because the region’s digital systems are increasingly interconnected. 

A weakness in one country’s information environment can affect public trust in government, businesses, and ASEAN communities. 

Based on UN principles, the best way to address AI’s global risks is to bring countries together to adopt shared principles and commitments on AI governance.

The need for interoperable governance

ASEAN does not require identical national laws to promote regional cooperation. 

However, it needs compatible principles and interoperable practices within the ASEAN AI governance framework.

In 2024, ASEAN Secretariat issued an ASEAN Guide on AI Governance and Ethics, and actively being updated in 2025 and 2026. 

This document serves as practical guide for government and organisations in the region that wish to design, develop, and deploy traditional AI technologies in commercial and non-military or dual use applications. 

This guide focuses on encouraging alignment within ASEAN and fostering the interoperability of AI frameworks across jurisdictions. 

It also includes recommendations on national level and regional level initiatives that governments in the region can consider implementing to design, develop, and deploy AI systems responsibly. 

Fortunately, ASEAN has representative that served as Co-Chair of Independent International Scientific Panel on AI, Maria Ressa, and representative that served as founding member of newly launched (during 81st UNGA, Digital@UN) AI for Good Global Commission, Singapore’s Minister for Digital Development and Information Josephine Teo.

Priority areas for ASEAN interoperability

Five areas deserve priority to reach interoperability between ASEAN. 

First, ASEAN governments should develop a shared vocabulary for AI risk; second, cross-border data protection; third, ASEAN should promote common approaches to content labeling, provenance, verification, and rapid cross-border response; fourth, shared testing methods, evaluation benchmarks, and professional training; last but not least, capacity building.

To strengthen international cooperation on AI capacity building, 29 countries, including five from ASEAN, became founding members of the World AI Cooperation Organisation (WAICO) at the 2026 World AI Conference and High-Level Meeting on Global AI Governance in July.

The meeting also producing commitments from China by providing 5,000 places in AI training and seminar programs for developing countries over the next five years; develop international AI application cooperation centers with ASEAN, the League of Arab States, the African Union, CELAC, the Shanghai Cooperation Organisation, and BRICS; and enable 30 countries to deploy the MAZU AI-powered meteorological warning system. 

These are forward-looking policy commitments whose execution will need to be verified through subsequent projects and public reporting.

Regulatory sandboxes for ASEAN innovation

A regulatory sandbox can help ASEAN manage the tension between innovation and precaution. 
Under a sandbox model, selected AI applications are tested within a controlled environment, with defined objectives, monitoring requirements, safety controls, and exit conditions.

A regional sandbox network could allow governments and innovators to share lessons without requiring immediate, full-scale deployment. 

It could also help identify which safeguards are effective across different legal and cultural environments.

However, a sandbox should not become a way to avoid accountability. Sandbox participation should not remove responsibility for deliberate misconduct, major harm, or violations of individual rights.

Human control must remain central

As AI agents become more autonomous, human control becomes increasingly important. 

AI systems should have unique identities, limited permissions, monitored tool access, and clearly defined operational boundaries. 

High-risk actions should require human approval, while logs should record system behavior and decision points.

For ASEAN governments, this principle is especially relevant to public sector applications. 

An AI system may help identify service gaps, summarise public feedback, or support policy analysis. But final responsibility must remain with accountable public institutions and authorised officials.

Human control also means preserving the right to challenge or correct an AI-assisted decision. 
Citizens should be able to understand when AI has influenced a decision, request clarification, and seek human review. 

If Elon Musk said to The Economist that AI could exceed the sum of all human intelligence by around 2031, we must make human control remain central in that scheme. 

We don’t want "the bad things” that OpenAI’s Sam Altman said the world should accept from AI in his discussion with Politico recently to ever happen.

-------------------------------
 

The author is a PhD Student of Communication and Global Development at School of Journalism and Communication Tsinghua University. She previously led Marketing,  Communication and Customer Operation at the Health Transformation and Digitalization Team (GovTech Health), Indonesian Ministry of Health.