Beyond the Guide: Navigating the Next Phase of Artificial Intelligence Governance in ASEAN

The newly released ASEAN Guide on AI Governance and Ethics (ASEAN AI Guide) is a positive development, yet its long term impact hinges on what comes next. As artificial intelligence (AI) rapidly advances, ASEAN must balance voluntary guidelines with targeted, binding measures to address the core challenges faced by its diverse member states (AMS). Balancing innovation and security while bridging the digital divide in a fast-evolving and uncertain strategic landscape will not be an easy feat. However, with a proactive, collaborative, and targeted approach, ASEAN can mitigate AI risks while harnessing this technology’s potential for rapid, inclusive, and sustainable growth.

The ASEAN AI Guide is guided by core principles such as transparency, fairness, human-centricity, accountability, and robustness. Its practical recommendations spanning the AI lifecycle from design to deployment provide a solid foundation for organisations and businesses to align practices and build public trust. Furthermore, the Guide’s risk-based approach to determining the necessary level of human oversight for different AI decision-making processes is highly appropriate, enabling significant flexibility while still mitigating the most potent risks. The national-level recommendations to nurture talent, support innovation, invest in research and development (R&D), and raise citizen awareness are also timely and comprehensive.

The region-wide voluntary AI guidance is therefore a crucial first step. Hitherto, AI governance in ASEAN has primarily occurred at the national level; regional initiatives such as the ASEAN Framework on Digital Data Governance do not address AI as their primary focus. Nevertheless, the true measure of the ASEAN AI Guide’s success will lie in its implementation across the region. The ASEAN AI Guide alone may prove insufficient to tackle the complex and multifaceted challenges of governing AI in ASEAN.

The window of opportunity to shape AI governance in Southeast Asia may be narrower than expected.

ASEAN’s unique context presents complex hurdles for AI governance. Promoting innovation while ensuring security is difficult given the region’s varied levels of digital maturity. While Singapore is clearly ahead of the rest, other AMS are still in the early stages of their AI journeys. The Guide rightly emphasises interoperability but more work is needed to harmonise policies and create a level playing field for AI adoption.

The voluntary guide rightly notes that AI has the potential to dramatically increase productivity. Yet this technology risks displacing humans from a wide range of jobs. Equipping ASEAN’s workforce with the necessary skills to thrive in an AI-driven economy is crucial for inclusive growth and social stability.

The current global AI governance context adds another layer of complexity for ASEAN. The intensifying strategic competition between the major powers and the EU AI Act’s implementation threatens to create a fragmented technological landscape, potentially exerting conflicting pressures on ASEAN’s approach to AI governance. Navigating these strategic complexities requires ASEAN to learn best practices while developing its own approach. Effective regional coordination and targeted diplomacy will be essential to secure the region’s interests and reflect its values in emerging international norms on AI.

Looking to the next phase, ASEAN should first promptly establish the AI Governance Working Group proposed in the Guide, with each AMS appointing an AI ambassador. This will create an empowered forum to drive implementation, monitor progress, share best practices, and interface with dialogue partners. More importantly, these ambassadors can champion AI governance in their national contexts and represent domestic concerns regionally.

AMS should also initiate discussions on an ASEAN binding treaty or code of conduct for high-risk AI applications, particularly those with military implications. While agreeing on legally binding measures will be far more challenging than agreeing on voluntary guidance, it is a necessary step to mitigate AI’s existential risks and prevent its misuse.

One key concern is the development and adoption of autonomous lethal weapons systems, which can select and engage targets without meaningful human control. The proliferation of such systems could lead to devastating consequences, including the unintended outbreak or escalation of conflicts. The agreement should also address other high-risk AI use cases that could harm individuals or society and cover issues like data governance, algorithmic bias, and the use of frontier AI models in critical decision-making processes. By setting clear and binding guidelines and boundaries, ASEAN can foster trust in AI technologies and ensure their responsible deployment.

Last, ASEAN should launch a region-wide effort to upskill the workforce on AI literacy and human-AI interaction. With 60 per cent of the region’s population under 35 years of age and a generally optimistic outlook among ASEAN youth regarding the impact of technology on jobs, equipping young people with the necessary skills to effectively interact with AI systems is vital. School curricula should go beyond technical skills to include AI ethics, governance, and sector-specific applications. In this regard, the region could look toward countries with more advanced AI approaches to access valuable training resources and deepen its understanding of AI applications. Moving forward, ASEAN could explore establishing AI training exchange programmes with its Dialogue Partners to facilitate knowledge sharing and capacity building.

The window of opportunity to shape AI governance in Southeast Asia may be narrower than expected. ASEAN must act decisively and swiftly to shape the trajectory of AI development and deployment in the region and beyond. With determination and a shared vision, ASEAN can hopefully rise to the challenge and leverage AI to unlock new frontiers of growth and innovation while ensuring comprehensive security for all.

2024/121