Thailand’s proposed artificial intelligence (AI) law is facing scrutiny from industry representatives who argue it could stifle local innovation if it doesn’t adequately support startups. While the draft aims to establish rules covering ethical considerations, consumer protection, and accountability, concerns remain that it may create unnecessary barriers for nascent AI companies. The Thai government is currently revising the draft legislation, which will be submitted to the Digital Economy and Society Ministry and the Office of the Council of State for further consideration.
Navigating the AI Landscape in Thailand
Thailand’s AI market is estimated to be worth around $50 billion, but a significant portion of this revenue is captured by foreign technology providers, with local companies earning less than $300 million. This imbalance is partly attributed to the challenges faced by domestic startups in securing funding and competing with established international players. “Local startups often see little incentive to choose Thai AI services over established foreign alternatives,” noted Khanchai Chawansathit, president of the AI Entrepreneurs Association of Thailand (AIEAT). He added that the new law should reflect Thailand’s economic realities rather than solely focusing on controlling risks.
The AIEAT estimates that Thailand has between 900 and 1,000 AI-focused enterprises, including early-stage and informal operations. However, only 176 new AI-related companies have been registered in Thailand over the past 2.5 years. This slow emergence of new AI work is concerning, especially as AI is expected to create new jobs and disrupt existing ones. Around 300,000 graduates and more than half of job seekers are struggling to find employment in a difficult job market.
Key Concerns for Local AI Firms
Industry representatives are calling for the AI law to include clearer liability structures, practical support for local firms, and robust protection of intellectual property (IP). The current draft, while containing potentially useful mechanisms, may impose disproportionate regulatory burdens on small businesses. “We are not saying we don’t want laws, but we want the law to understand our reality,” said Saphathana Nopakun, a committee member of the Thai Startup Association. She explained that startups typically have small teams with limited resources and legal expertise, making it difficult to navigate complex regulations.
Liability and Risk Classification
A significant point of contention is the draft’s broad definition of an AI system provider, which groups together various actors like system integrators, downstream application developers, and foundational model providers. This can complicate the determination of responsibility when an AI system goes wrong. Saphathana proposed a shared liability model, assigning obligations based on each participant’s level of control and reasonable access.
The draft also outlines regulatory intervention for applications affecting sensitive sectors such as state security, health, energy, and transport. However, developers may be left uncertain about whether their application qualifies as “high-risk.” The draft suggests classifying applications based on factors like the scale of impact, system autonomy, reversibility, affected groups, and the effectiveness of human oversight, rather than broad sector-wide categories. This approach, Saphathana noted, could provide greater certainty.
Intellectual Property Protection
The industry is also concerned about IP protection. The draft allows authorities to seek technical documents, training data, and related information if an AI incident is serious. “These are things that could constitute our entire IP, which could lead to decisions about whether it’s even worth operating in Thailand,” Saphathana stated. Foreign foundational model providers might respond by delaying new model releases, restricting access, or imposing regional restrictions, potentially making their services more expensive or unavailable to Thai users.
Support for Startups
The draft law is seen as falling short in promoting AI development, while existing regulations could potentially oversee AI within their current scope. The Thai government needs to design a framework that fosters innovation and competitiveness, enabling Thai AI-driven services to be exported. “The objective is not just for growth, but to strengthen Thailand’s competitiveness and develop AI-driven services that can be commercially exported,” said Ploy Chanchareon, deputy executive director of the Electronic Transactions Development Agency (ETDA).
For startups, the core value propositions are cost reduction and efficiency increases. However, AI’s ability to operate with a degree of automation and generate outcomes not anticipated by humans differs from conventional technologies. This unpredictability can discourage businesses from adopting AI due to potential liability concerns. The ETDA emphasizes a risk-based approach, aiming to assign obligations according to potential risk and damage, and to promote fair enforcement and regulatory coherence, including regulatory sandboxes.
Proposed Solutions and Future Directions
Saphathana proposed several potential safeguards, including access to computing infrastructure, training datasets, and secure review environments. Other suggestions included qualified independent experts, confidentiality protocols, and reduced fees for certification and fast-track sandbox access. She stressed that startups need mechanisms that allow them to handle compliance without excessive costs or resources.
For high-risk AI, the draft suggests that operators would first be required to correct non-compliant instances before penalties are imposed. This risk-based framework aims to fill gaps in existing laws where necessary and promote regulatory coherence. The proposed approach would involve investigations based on evidence, starting with system logs and incident reports, and requiring companies to maintain access logs and data retention limits.
Nuttanong Choomphong, president of the ETDA, highlighted that Thailand needs clear operational rules alongside its AI law. He noted that regulators should examine international regulatory developments while designing a framework appropriate for Thailand’s specific circumstances. “The objective is not to stifle innovation, but to foster it,” he stated.
Are Varee, vice president of the Council of State, called for support funds and plans for workers displaced by automation, while also cautioning against excessive AI hype. He advocated for finding startups that can solve specific national problems, such as improving the efficiency of the 30-baht healthcare scheme or increasing transparency in medical procurement, rather than solely focusing on nurturing new startups.
The debate underscores the critical need for Thailand’s AI law to strike a delicate balance. It must provide necessary safeguards and accountability without hindering the growth and competitiveness of its domestic AI industry, ensuring that the nation can harness the full potential of artificial intelligence.
