What’s Covered?
The document offers ASEAN governments a shared policy playbook for addressing the fast-growing impact of Gen AI tools, from chatbots and image generators to large-scale foundation models. It extends the general ASEAN AI Guide (2024) by adding layers specific to Gen AI systems and their content-generating capabilities. These include risks tied to privacy, intellectual property, misinformation, impersonation, and embedded social biases.
It outlines six core risks that Gen AI may amplify:
- Mistakes and anthropomorphism
- Inaccurate content and disinformation
- Deepfakes, impersonation, and malicious use
- IP rights infringement
- Privacy breaches
- Biased outputs
It also flags frontier and systemic risks—like model autonomy or misuse by bad actors—which are not widespread in the region yet but merit longer-term monitoring.
Nine policy areas are proposed to manage these risks:
- Accountability – assigning responsibility throughout the AI lifecycle
- Data – ensuring lawful collection, privacy safeguards, and traceability
- Trusted Development and Deployment – promoting safety, explainability, and proportional governance
- Incident Reporting – encouraging systems to log and escalate failures or misuse
- Testing and Assurance – establishing model evaluation practices
- Security – mitigating attacks such as data poisoning or prompt injection
- Content Provenance – embedding markers to trace origin of AI-generated content
- Safety & Alignment R&D – supporting research into AI alignment, robustness, and misuse prevention
- AI for Public Good – encouraging regional collaboration and ethical innovation
To ground these recommendations, the Guide presents four ASEAN-based use cases:
- PhoGPT (VinAI, Vietnam)
- Project Moonshot (AI Verify Foundation, Singapore)
- Responsible AI Program (Accenture, ASEAN-wide)
- ThaiLLM (NSTDA, BDI, VISTEC, Thailand)
The Guide frames Gen AI as both a major economic opportunity (estimated S$6 trillion across APAC by 2038) and a significant policy challenge that cuts across privacy, security, and ethical innovation. Its strength lies in integrating foundational AI principles with new Gen AI-specific risks, such as foundation model opacity and scaling harms.
💡 Why it matters?
ASEAN member states are under pressure to keep pace with Gen AI without sacrificing trust or regional values. This guide offers a harmonized approach—practical, forward-looking, and aligned with global benchmarks—to help ASEAN regulators and developers share responsibility. It makes clear that building a safe Gen AI ecosystem isn’t just about rules, but also about enabling public-sector innovation, collaborative oversight, and cross-border policy alignment.
What’s Missing?
While the guide offers solid principles and use case examples, it stays high-level in terms of enforcement tools. There’s little practical detail on how individual ASEAN regulators might audit, penalize, or intervene in Gen AI misuse—especially when models are developed outside their jurisdiction. The treatment of General Purpose AI and foundational model accountability is also somewhat cursory. The absence of clear thresholds for risk classification or certification could slow down adoption of shared standards.
There’s also room for more detail on community participation and non-state actor involvement in Gen AI governance, especially in highly decentralized or informal economies within ASEAN.
Best For:
Regional policymakers, government technology offices, AI developers working in Southeast Asia, and regulators trying to align domestic law with international Gen AI standards. Also helpful for international actors (like multilaterals or NGOs) advising ASEAN on AI governance capacity-building.
Source Details:
Title: Expanded ASEAN Guide on AI Governance and Ethics – Generative AI
Issued By: ASEAN Member States, building on the ASEAN Guide on AI Governance and Ethics (2024)
Expertise: Written by regional AI governance experts and public agencies, informed by regional practice and global standards. The guide reflects consensus-building across diverse legal and technological contexts. It includes policy recommendations grounded in emerging use cases from Vietnam, Thailand, Singapore, and pan-ASEAN deployments.
Purpose: To provide a regionally aligned approach to Gen AI policy that is risk-aware but still innovation-friendly, and to serve as a living resource for national regulatory strategies.
