Southeast Asian countries want AI to support development, but dependence on foreign systems could narrow their choices when politics or security tensions rise. In a China-US Focus opinion piece, Sun Chenghao argues that stronger local expertise and regional cooperation can help governments use AI without handing over the judgement that should remain their own.
Watch Desk analysis
What happened
Chenghao, a fellow at Tsinghua University’s Center for International Security and Strategy, draws on a recent seminar in Hanoi to describe the dilemma: AI systems can rely on outside computing, model updates and technical support, while becoming embedded in public services and critical infrastructure. If access is disrupted, he argues, the consequences could reach beyond technology procurement and constrain policy choices.
He notes that countries are taking different approaches. Singapore is positioned to pursue research and help shape governance rules; Malaysia and Vietnam are emphasising digital infrastructure and industrial development; countries with weaker foundations still need basic skills and infrastructure. Chenghao argues that using foreign technology and building domestic capability need not be competing choices. Read Chenghao’s opinion piece.
Why it matters
The practical question is not simply which country supplies a model. It is whether governments can assess how a system works, check its outputs and suspend it when needed. Chenghao highlights this particularly for maritime surveillance, intelligence analysis and cyberdefence, where a system’s account of a threat could influence consequential decisions.
He argues that ASEAN could help members pool research, testing resources and training rather than making each country build everything alone. The piece also points to the ASEAN AI Safety Network under development and the ASEAN Responsible AI Roadmap for 2025–30 as potential foundations for regional cooperation.
Our read
This is an argument about keeping room to manoeuvre, not a claim that every foreign AI partnership is a trap. That distinction matters: countries may work with several partners, but they need enough expertise to understand the conditions attached and judge whether a tool suits their needs. “Strategic autonomy” can sound like conference-room wallpaper; the useful test is whether officials can verify a system’s output and stop using it.
What to watch
- Whether ASEAN’s proposed AI Safety Network develops practical arrangements for incident notification and joint response.
- Whether regional research, testing and training efforts expand beyond broad commitments.
- Whether governments build independent capacity to assess AI used in public services and security.
Discussion spark: Should Southeast Asian governments prioritise access to the most capable foreign AI systems, or invest more in local expertise and shared regional safeguards, even if that slows adoption?
Sources and evidence
- How Can Southeast Asia Strengthen Its Hand? – Sun Chenghao – CHINA US Focus (16 October 2026, 11:48 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.