ASEAN, explained.

A close-up view of a single cigarette (Arun Anoop/Unsplash)

Natural-language-processing tools monitoring social media can detect novel tobacco-marketing tactics, enabling regulators to act swiftly (Arun Anoop/Unsplash)

AI-driven tobacco control in ASEAN requires unified data standards

Aligning the diverse digital infrastructure across member states offers a chance for effective AI deployment to help the region’s 120 million smokers.

Mochammad Fadjar Wibowo,Muhammad Daniel Azlan Mahadzir ,Elyssa Liu

12 August 20254 min read

Tobacco use remains a leading cause of preventable death in Southeast Asia, home to over 120 million smokers. While AI can support smart efforts in controlling tobacco export and use, it is only possible if the region has a standardised data entry system that allows every country to effectively and legally share their data with each other. Yet, the diversity of data systems across the region, from Singapore’s fully digital health records to paper registries in emerging economies, poses both challenges and opportunities for working together.

All ASEAN member states, except Indonesia, have signed and ratified the WHO Framework Convention on Tobacco Control, but national statutes and data-privacy regimes vary widely.

Singapore combines its Tobacco (Control of Advertisements and Sale of Tobacco Products) Act with the Personal Data Protection Act (PDPA), enabling approved research on de-identified patient data via its National Electronic Health Record system.
Thailand enforces its Tobacco Products Control Act of 1992 and Customs Act to mandate health warnings and criminalises contraband, while piloting a centralised smoking-cessation registry under the Ministry of Public Health.
Malaysia routinely publishes excise-tax revenues and retailer-licensing data but still relies on fragmented, state-level cessation surveys; its proposed Control of Tobacco Products and Smoking Bill 2022 aims to harmonise these datasets.
The Philippines enforces its Data Privacy Act of 2012 for all personal-data processing and delegates tobacco-product oversight to the FDA, but cessation-programme outcomes are captured in separate, manually updated registries.
Indonesia, despite not being part of the WHO Framework Convention on tobacco control, still collects extensive metadata through its Health Information System legislation, yet most primary-care records remain siloed within district health offices, complicating nationwide AI deployment.
Cambodia, Laos and Timor-Leste, meanwhile, often lack digital health-information infrastructures altogether, relying instead on periodic WHO-led surveys for population-level prevalence data.

Harnessing AI across heterogeneous datasets

Predictive analytics thrive only when fed by high-volume, high-variety and high-velocity data. But the application of AI in tobacco control offers varied policy proscriptions.

When applied to customs and retail-sales data, this technology can spotlight emerging smuggling routes before they balloon into large-scale illicit trade. Machine-learning models trained on de-identified, multi-country patient cohorts can also personalise cessation programs by adapting motivational messages and pharmacotherapy regimens based on real-world outcomes.

A no smoking sign in red (Kristaps Solims/Unsplash)

Information sharing remains fragmented hence preventing sound decision making processes (Kristaps Solims/Unsplash)

Finally, natural-language-processing tools monitoring social media can detect novel tobacco-marketing tactics, enabling regulators to act swiftly. Pilot research in Indonesia has already shown that AI-enhanced mobile-health interventions can reduce cardiovascular-disease risk, which is an approach adaptable for smoking-cessation outcomes. Similar work has further demonstrated district-level capacity to develop and validate machine-learning solutions for health-behaviour monitoring. Emerging policy analyses, such as the “Generative AI Policy and Governance Considerations for Health Security in Southeast Asia,” advocate harmonised frameworks to govern AI deployment in public.

Governance gaps and the need for unified architecture

Legal mandates provide the authority to collect and share data, but without well-designed architecture and governance mechanisms, information sharing remains fragmented hence preventing sound decision making processes.

For example, Malaysia’s Ministry of Finance publishes monthly customs seizure reports, yet health agencies rarely incorporate these figures into cessation outreach planning. Meanwhile, Thailand’s inter-ministerial task forces align taxation with cessation services but have yet to feed compliance audits into a dashboard accessible to other ministries.

To break down these silos, ASEAN could adopt a common data-exchange architecture featuring a shared data model.

To realise fully AI-driven tobacco control, ASEAN leaders need to harmonise data-protection laws.

For instance, a Fast Healthcare Interoperability Resources (FHIR)-based system could ensure consistency in how smoking status, treatment outcomes and enforcement actions are represented. This would also ensure automated de-identified feeds between ministries of health, finance and customs, underpinned by clear legal protocols as already exemplified by Singapore’s PDPA exceptions for public-health research. A regional oversight body could also be empowered to audit privacy compliance and grant “safe harbour” accreditation to national systems meeting interoperability criteria.

Public trust in a regional roadmap

A unified data-sharing ecosystem must be balanced by strong privacy safeguards to maintain public confidence.

This involves strict de-identification standards to ensure individuals cannot be re-identified, even within federated analytics environments. It also requires independent oversight committees, comprising legal experts, ethicists and community representatives, to audit uses of shared data and AI algorithms. Transparent, periodic reporting on how aggregated data and AI-derived insights inform policy will build legitimacy, reinforcing the social licence for data collection and fostering broader community engagement.

To realise fully AI-driven tobacco control, ASEAN leaders need to harmonise data-protection laws. There is a need to invest in core digital infrastructure so that even lower-middle-income and low-income countries can capture and share basic cessation and enforcement metrics. They must also develop common technical standards and codes to underpin a regional data hub, and launch capacity-building initiatives pairing mature systems (such as Singapore, Thailand) with developing ones (Cambodia, Laos).

By acknowledging the region’s diversity of data maturity and committing to a harmonised, interoperable framework, ASEAN can move from fragmented national efforts to a united, predictive strategy, accelerating progress toward a smoke-free future.

About the authors

Mochammad Fadjar Wibowo

Mochammad Fadjar Wibowo is a global health policy researcher focusing on the evaluation of digital health interventions and governance.

Muhammad Daniel Azlan Mahadzir

Muhammad Daniel Azlan Mahadzir is an award-winning clinical nutritionist and public health scientist based in Malaysia and Singapore.

Elyssa Liu

Elyssa Liu is a global health lawyer specialising in health emergencies and outbreak preparedness.

Topics

ASEAN Cyber & technology