Southeast Asia needs AI exit options, not AI independence

Summary

Writing in Fulcrum, the ISEAS – Yusof Ishak Institute's publication, on 5 October 2026, Zenobia Chan argues that Southeast Asian governments should abandon the goal of full AI independence. Chan teaches at Georgetown University and is a visiting fellow at ISEAS. In her view, most countries in the region cannot replicate the entire stack, from semiconductors and data centres to frontier models and software tools. Chasing that goal wastes resources and distracts from achievable aims. Instead she proposes AI resilience: the capacity to manage dependence on foreign technology through diversification and reversibility.

Chan rests resilience on three pillars. The first is keeping the option to switch providers alive. Governments should map dependencies at every layer of the stack, monitor exclusivity agreements, secure data access and export formats in contracts, and test portability before buying. The second is making systems work in the region's languages and conditions. Rather than each country training its own foundation model, that means shared datasets, benchmarks and local adaptation skills. The third is service continuity: small, locally runnable models in critical areas such as health, finance and transport that can step in if a foreign service is cut off. For data centre investments, she calls for scrutiny of ultimate ownership, operational control and which foreign laws apply. She suggests ASEAN develop shared language datasets, testing protocols, incident reporting systems and procurement templates.