World Bank report highlights Thailand’s AI opportunities and gaps

The World Bank’s World Development Report 2026: The Promise of Artificial Intelligence assesses the impact of artificial intelligence, or AI, on about 6.8 billion people in low- and middle-income countries.

It identifies AI as a potentially important tool for helping developing countries overcome longstanding constraints in staffing, knowledge and the quality of public services.

However, the benefits of AI will not materialise automatically.

Countries that lack electricity systems, internet access, workforce skills, databases and strong regulatory institutions may benefit less than developed countries and risk facing an even wider productivity gap.

For Thailand, the report highlights the country’s role in several areas.

Thailand is among the developing countries with a high export value for goods that support AI systems, is among the countries attracting high levels of data-centre investment and was included in the World Bank’s survey of business AI use, which covered 360 Thai companies.

AI spreading faster than earlier innovations

The World Bank classifies AI as a “general-purpose technology”, alongside the steam engine, electricity, computers and the internet, because it has the potential to restructure not only individual industries but also entire systems of production, work and public services.

The crucial difference is that AI is spreading much faster than earlier innovations.

It took about 80 years for the steam engine to reach low-income countries, around 40 years for electricity and about 20 years for the internet.

Just six months after ChatGPT was launched, however, middle-income countries already accounted for about half of its global use.

The report analyses AI through three factors: technological capabilities, the concentration of power and domestic complements.

AI can perform complex cognitive tasks, such as analysing medical conditions, forecasting the weather and advising farmers, but the upstream technologies, including chips, data centres and advanced models, are controlled by a small number of companies in only a few countries.

For Thailand, an electronics industry base and investment in data centres represent only part of its readiness.

Generating economic value from AI will also depend on workforce skills, Thai-language data, businesses’ capacity to apply the technology and the ability to connect infrastructure with activities that create added value within the country.

No need to compete in building large models

A key approach proposed by the World Bank is “Adopt, Adapt and Advance”: countries should begin by adopting existing technology, then adapt it to the local context and develop advanced technology of their own when ready.

Developing frontier AI models requires investment, chips, data centres, large datasets and world-class researchers.

In 2026, major US AI companies and cloud providers, Alphabet, Amazon, Meta, Microsoft and Oracle, plan a combined investment of about US$775 billion, or approximately THB25.19 trillion.

That sum exceeds the size of many national economies.

A chart in the report compares it with Thailand’s 2025 gross domestic product, estimated at US$577 billion, or approximately THB18.75 trillion.

The five technology companies’ planned investment is therefore about 1.34 times Thailand’s GDP.

The comparison indicates that attempting to compete in developing large models from the outset may not be cost-effective for Thailand and most developing countries because they would be competing against companies whose resources exceed the economic output of many countries.

The World Bank therefore recommends that these countries start with ready-made tools and adapt them to local languages, laws, cultures and working practices.

It also recommends developing “Small AI”, or smaller models designed for specific tasks, which can run on ordinary devices or work offline without requiring large data centres.

For Thailand, this approach means prioritising the adaptation of AI to support the Thai language, documents, databases and industry-specific knowledge over seeking to build large models spanning every part of the technology value chain.

4.5% of jobs face AI replacement risk while 16.2% may benefit

The effect on labour markets in developing countries may differ from that in wealthy economies.

The report estimates that about 4.5% of jobs in low- and middle-income countries involve tasks that generative AI could perform instead of workers, compared with 14.2% in high-income countries.

Conversely, about 16.2% of jobs in low- and middle-income countries are likely to benefit from AI as a tool that enhances workers’ capabilities rather than replacing them entirely.

The jobs facing the earliest risks include those in business services, call centres, office support, finance and software, as well as entry-level positions focused on processing information or producing content.

Most workers in agriculture, retail, hotels and small businesses, however, still perform jobs requiring physical activity and interaction with customers.

The figures of 4.5% and 16.2% are averages for low- and middle-income countries, not a direct assessment of Thailand’s labour market.

They should therefore be used to indicate the direction of the effects rather than to conclude that the same proportion of jobs in Thailand will be replaced.

The World Bank also surveyed companies in India, Jordan, Kenya, Mexico, Nigeria and Thailand.

It found that about one in five small businesses already used AI chatbots in their operations, compared with about one in four in the United States.

The survey included 360 Thai companies.

The participating companies were manufacturing, construction and service businesses with at least five employees.

Their uses of AI ranged from finding and summarising information, writing and editing text, translating and communicating with customers to analysing data, controlling production processes and developing software.

However, the published findings are combined averages for the participating developing countries, not results specific to Thailand.

They should therefore not be cited as the rate of AI use among Thai businesses.

Entrepreneurs in the developing countries covered by the survey expect that AI could help raise productivity by about 11% over the next three years under the most likely scenario, while employee numbers are expected to increase by around 7%.

This indicates that businesses still view AI more as a tool for expanding output than solely as a means of reducing headcount.

In Thailand’s context, the key issue is therefore not merely to train workers to use AI tools, but also to redesign jobs, management processes and the division of tasks between people and technology so that AI delivers genuine productivity gains.

Thailand has opportunities in the AI hardware chain

Beyond adopting AI, Thailand is identified as one of the developing countries with a role in the AI hardware supply chain.

The report states that goods supporting AI development increased their share of global trade from about 13% in 2023 to nearly 17% by the end of 2025, alongside rising investment across the semiconductor, electronics and data-centre equipment supply chains.

Low- and middle-income countries account for about one-third of the export value of these goods.

Under a narrower definition of AI-related goods, their share is about 39%.

The developing countries with the highest export values for goods supporting AI systems in 2025 were China, Mexico, Malaysia, Vietnam and Thailand, in that order.

Thailand therefore ranked fifth among developing countries in the report.

The data indicate that Thailand is not starting from zero in the AI economy, but already has a production base within the value chain, spanning electronics, semiconductors and data-centre equipment.

However, a hardware manufacturing base alone may not be enough.

The report notes that value in the AI chain remains concentrated among a small number of companies.

Thailand’s opportunity will therefore depend on its ability to move from contract manufacturing and exports into design, software, cloud services, data services and higher-value AI applications.

Thailand ranks in the top 10 for data-centre investment

The report also states that data centres accounted for more than one-fifth of global greenfield foreign direct investment in 2025, making them one of the leading activities attracting investment in new projects.

Although developed economies remained the main destinations, four middle-income countries were among the world’s 10 largest recipients of data-centre investment: Brazil, Thailand, India and Malaysia.

This position allows Thailand to benefit from foreign investment, economic growth and additional infrastructure for data processing and AI.

However, the World Bank observes that the number of jobs data centres can generate remains a matter of debate.

They also create environmental externalities through their heavy consumption of electricity, water and other resources.

The benefits of data centres to Thailand should therefore not be measured solely by investment value, but also by the extent to which that investment can forge links with domestic entrepreneurs, software developers, workers and the country’s technology ecosystem.

AI can improve public services if data are ready

The report gives examples of AI use in developing countries, including a weather-forecasting system in India’s Telangana state that has helped small-scale farmers adjust production plans and save up to about US$560 per person, or approximately THB18,200 per person.

In Kenya, an AI system is used to assign more than 10,000 cases a year to over 1,500 mediators, helping to reduce case backlogs.

Bangladesh, meanwhile, uses AI to screen medical images, increasing the number of patients checked for diabetes-related eye complications by nearly 40% a day.

The main constraints facing public agencies vary.

Citizen-facing services are often hindered by officials’ skills, internet connectivity and systems that do not support local languages.

Back-office functions face problems with the quality and availability of data, much of which remains scattered across paper records, legacy computer systems and knowledge held by individual officials.

In Thailand’s context, the report’s framework suggests that public-sector AI adoption should not begin solely with buying systems.

Agencies must also assess whether Thai-language and agency data are in formats the systems can use, whether officials can adequately check the results and whether agencies can maintain, improve and evaluate the systems continuously.

Governments must perform three roles

The World Bank proposes that governments perform three roles in relation to AI simultaneously: enabler, user and regulator.

As an enabler, the state must invest in electricity, broadband, education, digital skills, computing systems and local-language data, while helping small businesses access information on which types of AI are suitable for their operations.

As a user, the state should prepare data for AI, invest in staff skills and shift procurement away from one-off project purchases towards continuous testing, evaluation and improvement.

It must also retain rights of access to data and reduce dependence on any single vendor.

As a regulator, the state should begin by reviewing existing laws and regulations, use voluntary industry standards as a starting point and manage risks involving data protection, consumer protection, discrimination, safety, fraud and cybercrime.

For Thailand, the three roles are interconnected because the country has a hardware manufacturing base, data-centre investment and businesses that are beginning to adopt AI.

Fully realising the benefits, however, will still depend on human resources, Thai-language data, access to technology for small businesses and regulatory capacity.

Thailand’s challenge is not to build the largest AI model

Based on the report, Thailand has at least three links to the AI economy.

First, it is one of the five developing countries with the highest export values for goods supporting AI systems.

Second, it is among the world’s 10 largest recipients of data-centre investment.

Third, Thailand was included in the World Bank’s survey of business AI use, which covered 360 Thai companies.

These findings indicate that Thailand has a manufacturing base, infrastructure and a business-user sector, but the report does not assess how much value the country can capture from each area.

Thailand’s challenge should therefore extend beyond attracting data centres, increasing equipment exports or competing to build large models.

It must connect these foundations with higher business productivity, the development of AI software and services, Thai-language data and the adaptation of technology to solve the country’s problems in practice.