NESDC says 8.7m Thai workers are exposed to GenAI’s impact on jobs

Danucha Pichayanan, secretary-general of the National Economic and Social Development Council (NESDC), said on Monday (May 25, 2026), regarding the first-quarter 2026 Thai social situation report, “Crisis and Opportunity: The Future of Thailand’s Labour Market Amid the Arrival of AI”, that the NESDC had analysed data from the Labour Force Survey (LFS) for the fourth quarter of 2025 together with an assessment framework for occupations at risk of being affected by Generative AI (GenAI), based on the study by Gmyrek et al.

Of Thailand’s total labour force of 40.1 million, about 8.7 million workers, or 21.8% of the total, are in categories likely to be affected by GenAI.

When divided by the nature of AI’s impact, the findings are as follows:

Group at risk of task replacement (Task Replacement)

This group numbers 2.2 million, or 5.4% of all workers. It covers workers whose tasks AI is likely to replace, from partial substitution to the replacement of entire jobs, especially work that is repetitive and follows fixed patterns.

Most workers in this group have a bachelor’s degree or higher (55.8%), with the largest share graduating in business administration (41.8%). Their average income is THB27,820 per month, their average age is 36.5, and 53.0% work in Bangkok and surrounding provinces.

When classified by workers’ skill levels, only two groups were found to be at risk of replacement:

• Medium-skilled workers number as many as 1.9 million. The largest occupational group is clerks, secretaries and office support workers, accounting for 52.5%, including as many as 750,000 clerks. This is followed by customer service, reception and public relations work at 28.4%, including 310,000 contact centre staff, as well as accounting, finance and statistics staff at 13.7%.
• High-skilled workers number 330,000, especially marketing workers in advertising and marketing specialist positions at 31.1%, financial analysis and investment advisory roles at 24.7%, and information technology roles at 23.7%, such as software developers, database designers and computer network administrators.

A key reason these occupations are classified as at risk is that skills once considered human strengths and advantages, such as calculation, rules-based analysis and programming, have become tasks that AI systems can process more quickly, accurately and efficiently.

Group using AI to support work (Task Augmentation)

This group numbers 6.5 million, or 16.3% of all workers. AI can replace some parts of their work, but most core tasks still require human skills. For this group, the use of AI is therefore focused mainly on facilitating work and increasing efficiency.

Most workers in this group have below bachelor’s degree qualifications (72.7%), with an average income of THB21,506 per month and an average age of 42.3. They mainly work in Bangkok and surrounding provinces (37.3%), while the remainder are distributed across other regions in similar proportions.

When classified by skill level, most are medium-skilled workers, numbering 5.9 million, comprising:

• Sales staff (47.2%), who can apply AI to help recommend products, manage inventory and conduct preliminary analysis of customer behaviour, while still requiring people to negotiate and observe customers’ emotions and real needs.
• Shop owners (20.1%), who are beginning to use GenAI to manage businesses, forecast demand, plan orders in advance and use AI chatbot systems to answer questions and serve customers.
• Car, taxi and van drivers (17.2%), who use AI to help plan routes and navigate.

High-skilled workers number about 600,000, comprising (1) managers and executives (53.7%), (2) management and personnel roles (13.5%), and (3) public relations roles (9.1%). These workers have specialised expertise and make decisions that affect organisations, allowing them to use GenAI to reduce structured workloads with clear procedures, such as data gathering, document summarisation and in-depth analysis.

However, communication, interpersonal and cross-functional management skills remain core abilities that AI finds difficult to replace. Low-skilled workers number around 67,000, divided into delivery workers (82.0%) and meter readers (18.0%). Although these are basic-skill jobs, the use of technology to improve work efficiency is already beginning to appear.

An example from a study in the United States found that delivery workers used AI to process traffic conditions, weather, vehicle efficiency and parcel priority in order to improve transport planning for maximum effectiveness.

Challenges facing Thailand’s labour market

The situation above reflects that AI is not merely a technology of the future but is already creating impacts and risks for the structure of Thailand’s labour market, especially among medium- to high-skilled workers, as follows:

• Reduced employment in some groups and fewer work opportunities for new graduates. The labour market is facing challenges from organisational restructuring, with many industries likely to reduce employment, especially in jobs involving repetitive tasks or basic data analysis. Most of these are entry-level positions in organisations, creating challenges for new graduates, who may find it harder to enter the labour market than before. Many organisations can use AI to handle basic work and shift their focus towards recruiting people with high-level specialised skills and work experience who can command AI immediately.
• Labour movement from high-skilled jobs to lower-skilled jobs (Downshifting). As some high- and medium-skilled work can be replaced by AI, workers in these groups may adapt by moving from high- or medium-skilled jobs into lower-skilled work or jobs below their qualifications, because some low-skilled work is less affected by AI, such as switching to app-based hired driving or rider work.
• Technological progress towards Agentic AI and Physical AI. Rapid technological progress is set to broaden and clarify the scope of its impact. The full arrival of the Agentic AI era means that jobs with fixed procedures or repetitive logic, such as data analysis or clearly structured professional work, will see their roles reduced quickly. In the next phase, technology is also likely to be integrated with robots and autonomous systems, known as “Physical AI”. This will shift the impact from competition over “data-based work” to “physical and practical-skill work”, inevitably affecting employment directly in manufacturing, logistics and services.
• The speed of producing human resources and upgrading skills to keep pace with technology. Although the public sector recognises the challenge and has driven work through the National AI Action Plan (2022–2027) under the “MHESI for AI” policy, with a target of producing 30,000 AI personnel within three years, covering system development researchers, application engineers and general professional users, the key issue is the speed of implementation. Thailand faces an urgent need to upskill and reskill workers to shift their roles from operators to “AI managers”, in line with technological development and the rapidly changing needs of the business sector.

How Thailand should prepare for the AI era

Under these dynamics of change, Thailand needs to prepare for the impact of AI in several key areas, including:

• Accelerating broad AI adoption and upgrading human capital. The public sector should urgently support the application of AI to increase work efficiency among workers at all skill levels, because AI will become increasingly capable of replacing humans in more tasks and across a wider range of work. Creating an environment that allows workers to adapt and use technology appropriately is therefore urgent, as is developing sufficient data infrastructure and processing systems to meet future demand.
• Adjusting the role of high-skilled workers towards becoming “AI managers”. High-skilled workers need to strengthen their skills and AI literacy so they can supervise and assess AI outputs and operations effectively, especially in dealing with technical limitations such as inaccurate information, bias in datasets and data privacy protection.
• Developing legal frameworks and digital and AI governance, especially regulatory mechanisms and laws that support the protection of workers’ rights in hiring, performance assessment and dismissal processes where AI may be involved. This also includes defining legal liability in cases where AI systems cause damage, such as financial transactions, contract execution or automated operations without human supervision.

Laws should also be designed with flexibility so they can respond promptly to rapid changes in AI.