Thailand is taking a significant step toward establishing a comprehensive legal framework for artificial intelligence (AI) with the Electronic Transactions Development Agency (ETDA) opening public consultation on the principles of a draft of the country’s first AI Bill. The initiative reflects Thailand’s ambition to balance AI-driven innovation with ethical governance, public safety, and fundamental rights.
Following a two-year hiatus since the revised draft of AI Bill, which was presented on 2 May 2025, has been updated to reflect developments in both international best practices and domestic policy considerations. The ETDA will accept public comments on the principles of the draft of the AI Bill until 9 June 2025.
Key Features of the Draft Legislation
The draft of AI Bill law is built around five foundational pillars:
1. Risk-Based Regulation
Rather than prescribing a fixed list of high-risk or prohibited AI applications, the draft empowers sectoral regulators and a central enforcement agency to determine such classifications. This decentralised yet coordinated approach enables tailored risk governance in each industry.
High-risk AI providers will be required to:
- Implement recognized risk management frameworks (e.g., ISO/IEC42001:2023, NIST);
- Appoint a legal representative in Thailand (if based offshore);
- Report serious incidents to the enforcement agency.
High-risk AI deployers must ensure:
- Human oversight;
- Maintenance of operational logs;
- Data quality assurance;
- Notification to affected individuals where rights may be impacted.
2. Support for Innovation
The draft includes pro-innovation mechanisms, such as:
- Exceptions for non-commercial use of online data for AI development (e.g., text and data mining);
- The establishment of AI regulatory sandboxes where developers can test AI systems under controlled conditions. Participants acting in good faith will benefit from safe harbor protections, though civil liability for damages remains.
3. General Legal Principles
The draft incorporates key principles to guide AI governance, including:
- Non-discrimination in legal recognition of AI-generated decisions;
- Human accountability for AI-generated actions;
- Protection from unforeseeable AI behavior in contractual or administrative contexts;
- Rights to explanation and appeal, particularly where AI impacts individuals’ rights or interests.
4. Regulatory Oversight
Oversight will be carried out by the AI Governance Center (AIGC) under ETDA. The AIGC will be tasked with:
- Supporting AI policy implementation;
- Advising on ethical adoption;
- Operating AI sandboxes;
- Monitoring international developments and coordinating domestic readiness efforts.
5. Legal Enforcement
The enforcement framework authorizes regulators to:
- Order the suspension or removal of non-compliant AI services;
- Coordinate with internet service providers to block access to unlawful AI systems;
- Seize physical products containing prohibited AI if necessary.
The ETDA will review public comments and revise the draft accordingly, with an additional round of consultation expected. The proposed legislation is expected to feature prominently at the 3rd UNESCO Global Forum on the Ethics of AI, to be hosted in Bangkok later this month.