Thailand’s draft AI law: A new era for governance and innovation

Thailand is advancing its efforts to establish a national, comprehensive framework for artificial intelligence, with new draft legislation currently under consideration. Two draft instruments are already on the table: the 2022 Draft Royal Decree on Business Operations Using AI Systems (ONDE Draft Decree), prepared by the Office of the National Digital Economy and Society Commission and the 2023 Draft Act on the Promotion and Support for National AI Innovation (ETDA Draft Act), crafted by the Electronic Transaction Development Agency (ETDA).

In June 2025, regulators convened a public hearing to merge both texts into a single, more practical instrument that is currently being circulated as the Draft Principles of the AI Law (Draft Principles).

The original ONDE Draft Decree was proposed to ensure transparency, safety and fairness in commercial AI deployment. Mirroring the architecture of the EU Artificial Intelligence Act (EU AI Act), it introduces a tiered risk classification with outright bans of systems that manipulate human behaviour subliminally, perform social scoring, process sensitive attributes or run real-time biometric identification in public spaces.

By contrast, the ETDA Draft Act focuses on nurturing innovation: it calls for government-backed AI Innovation Testing Centres, promotes data sharing between the public and private sectors and offers contractual-risk templates to lower transaction costs for early-stage projects.

The Draft Principles seek to reconcile these two imperatives—guardrails and growth—under one statutory umbrella. The text affirms that AI is ultimately a human-controlled tool and enjoys no independent legal personality. It introduces two headline risk categories. “Prohibited-risk AI” covers systems whose harm cannot be mitigated; once designated, their use may be criminalised. “High-risk AI” may impact fundamental rights or public safety yet remains permissible subject to stringent duties. The Draft Principles aim to empower the competent authority or sectoral regulators to issue lists of Prohibited-risk AI and High-risk AI through subordinate legislation. 

The Draft Principles place a strong focus on individuals’ rights. Any person subject to an AI-driven decision must receive clear notice, an explanation of the underlying reasoning and a meaningful opportunity to contest the outcome. The Draft Principles further address the proliferation of AI-generated content, especially deepfakes and disinformation threatening national security or democratic processes—with the possibility of technical measures to trace and watermark synthetic media.

Innovation incentives have not been overlooked. Limited copyright exemptions for text and data mining, similar to the EU AI Act, will support model training, while an AI regulatory sandbox will allow companies to pilot novel applications with supervisory oversight. The Artificial Intelligence Governance Center, under ETDA, will conduct research and development on AI governance, provide advisory services to organisations adopting AI, support AI regulatory sandbox initiatives, compile national AI readiness statistics, monitor global trends in AI development and application and foster domestic and international cooperation networks.

Sectoral regulators will also be empowered to issue supplementary codes of practice on risk management, tailored to the specific needs of each sector. Although the Draft Principles are currently being reviewed, the Securities and Exchange Commission has already published the Regulatory Framework for the Use of Artificial Intelligence and Machine Learning in the Capital Market since 2023. Further, the Bank of Thailand and the Office of Insurance Commission held a public hearing this year on guidelines concerning the use of AI to mitigate sector-specific risks.

What comes next? ETDA is revising the consolidated Draft Principles following public consultation and may hold another public hearing—a process expected to run to late 2025. Businesses that design, import, sell or operate AI systems in Thailand should begin mapping their inventories, classifying use cases by risk and preparing governance documentation consistent with international best practices.

We would like to thank Borwornrut Soontharos for her contribution to this post.