TH-AI Passport faces whitewashing claim as ministry defends cost and benefits

Thailand’s TH-AI Passport project has become a new political test for the Digital Economy and Society Ministry, as officials defend the 1.621-billion-baht AI skills scheme while an opposition MP prepares to question whether its terms of reference were written in a way that favoured certain contractors.

The issue intensified after the ministry opened the TH-AI Passport Forum at 9.30am on June 11 to gather views and explain the project. The event was attended by Digital Economy and Society Minister Chaichanok Chidchob, Deputy DE Minister Boonthida Somchai, Deputy Interior Minister Jeseth Thaiseth, and Dr Karndee Leopairote, a Democrat Party list MP and deputy party leader.

The ministry presented the project as a national AI upskilling programme aimed at giving up to five million people access to generative AI tools, online learning, assessment, certificates, boot camps and user-management services.

However, People’s Party list MP Rukchanok Srinork, who chairs the House committee studying, preparing and monitoring the budget, said the forum appeared to be a “whitewashing” exercise designed to legitimise a project whose key conditions had already been fixed.

Rukchanok said Permanent Secretary for Digital Economy and Society Patchara Anuntasilpa had already confirmed that the TOR could no longer be changed because the project had entered the contract-management stage.

She argued that public consultation should have taken place before the project moved forward, not after contract details had already been locked in. She also claimed the first payment instalment had already been made before the consultation, although the contractor later said it had not yet received any money.

Rukchanok also raised concerns over the publicity section of the TOR. She said the requirements appeared to specify detailed conditions for digital screens or billboards nationwide, including in convenience stores and at Suvarnabhumi Airport, which only a small number of operators could provide.

She further alleged that the wording of the publicity section closely resembled the TOR for a National Credit Bank project under the Ministry of Higher Education, Science, Research and Innovation during the tenure of former minister Supamas Isarabhakdi, with only measurement units changed.

Rukchanok questioned whether the specifications were written to benefit connected business networks and restrict wider opportunities for Thailand’s IT industry.

She said that if public registration for the project begins, she will submit the matter to the National Anti-Corruption Commission immediately.

The House budget monitoring committee is scheduled to meet jointly with the House committee on legal affairs, justice and human rights on June 18. Rukchanok said the meeting would invite Chaichanok, the permanent secretary, officials involved in drafting the TOR, the NACC and the State Audit Office to provide explanations.

She urged the DE minister to attend in person, saying his presence would show sincerity and respect for Parliament and the public.

Seven questions behind the TH-AI Passport debate

The BangkokBizNews report also set out seven key questions and explanations surrounding the project.

The first question is why the government needs to buy AI services through an intermediary. The project’s explanation is that it is not only purchasing AI usage rights, but integrating more than 30 AI models from 14 providers, together with an online learning system, assessment, boot camps and user management for five million people. This requires a system integrator to connect all services into one platform.

The second question is whether the 1.621-billion-baht budget is too high. The ministry and project representatives said the average cost is about 27-28 baht per person per month, or around 330 baht per year. They said this covers Pro or Premium-level AI tools, training courses and skills-development activities, and is lower than many private-sector services.

The third question is why the project does not simply use free versions of ChatGPT, Gemini or other AI tools. Officials said the TOR requires generative AI tools at Pro or Premium level because free versions have limits on usage, deep analysis, repeated questions and advanced features, while the project is intended to support real use in study and work.

The fourth question is how success will be measured. The project aims to raise Thailand’s AI usage rate from 10.7% to 20% by 2027. Other indicators include actual user numbers, course completion, certificates issued and continued AI use.

The fifth question is whether the project can be cancelled or the TOR changed. Patchara said the project has already reached the contract-management stage, so the TOR cannot be amended because it forms part of the signed contract. However, details of implementation can still be adjusted within the contractual framework.

The sixth question is whether the contractor has already received payment. TH Joint Venture, the winning contractor, said it has not received any money from the project. Payment must follow work phases and delivery conditions set out in the contract.

The seventh question is whether the state must pay the full amount if actual usage is lower than expected. The matter is still under discussion. Officials said the state will follow the principle of “use as much as needed, pay only for actual use” based on active users, while the contractor said it has already incurred system-preparation costs but is willing to discuss an appropriate payment structure.

Ministry says Thailand needs wider AI access

Patchara said the TH-AI Passport project was designed because Thailand needs to upgrade digital and AI skills on a broad scale.

He said Thailand’s digital economy indicators, particularly those related to AI, had continued to decline, pushing the country below average and leaving it around 89th in global rankings. In ASEAN, he said Thailand was still not performing at a satisfactory level.

Patchara said the ministry had previously implemented targeted support programmes for specific groups, including technology-related groups, individuals and juristic persons, through its networks.

However, those programmes had not improved the overall picture enough. The ministry therefore designed a broader project to reach more people, in line with the government’s policy of upskilling, reskilling and reducing inequality.

He said the project would not simply hand out AI access and leave users to figure it out themselves. It would include awareness-building and engagement activities so people could understand what AI is and how to apply it in practical ways.

The target group is people aged over 15 and is divided into three main categories: students, government personnel and the general public, including small and medium-sized entrepreneurs.

Patchara said students were a key group because the gap between urban and regional areas is especially visible in education. Government personnel could use AI to reduce work time and improve the completeness of information. The general public and SMEs could use AI to support business and work opportunities, although SMEs are grouped under the general public because the ministry already has other mechanisms to support them.

He said the ministry also saw a clear gap in AI access between major cities and regional areas, not only between Bangkok and the provinces but also among large urban centres. The project is intended to give people outside major centres better access to AI tools and skills.

Budget comes from DE fund

Patchara said the project was created after the ministry considered using money from the DE fund, which had about 1.8-1.9 billion baht available outside the regular state budget.

He said the ministry assessed that around 50 million people remained economically active and could benefit from AI skills. It then set a minimum target of 10%, or about five million people.

The original ambition was to reach up to 10 million people, but the target had to be adjusted to match the available budget.

Although the money comes from an extra-budgetary fund, Patchara insisted that spending still had to comply fully with laws and government regulations, especially public procurement rules. The only difference, he said, is that money remaining in the fund at the end of a fiscal year does not have to be returned to the Treasury as state revenue.

He rejected criticism that the procurement process was limited to a 34-day invitation period. He said that period was only one part of the process, which had already passed through several committees, including the DE fund committee chaired by the prime minister or an assigned deputy prime minister.

Overall, he said the process took almost five months before the contract was signed.

Reference pricing and bidding process defended

Patchara also addressed concerns about reference pricing and whether companies that provided pricing information should have been allowed to bid.

He said the ministry had to seek price information from operators directly involved in the relevant technology and digital sectors because this was an IT and digital project. It could not collect realistic pricing from unrelated sectors such as construction.

He warned that if companies that provide reference pricing were banned from bidding, private firms might refuse to give pricing information to the state. If government agencies set prices by themselves, those prices might not reflect real market conditions.

According to Patchara, the ministry collected reference prices from eight companies. The bidding process attracted two groups and one individual bidder. One major bidding group had not been among the companies used for reference pricing and joined the bidding independently.

He acknowledged that the winning contractor included a company that had previously provided reference-pricing information, but said the process was open to the public and publicly announced at every stage. These included the procurement plan, public hearing and invitation to bid through ministry channels and the Comptroller General’s Department.

The project is divided into five delivery phases. Patchara said the first phase requires the contractor to submit both an implementation plan and preliminary work results, not just a few documents in exchange for payment.

He said the system must be tested, trialled with selected users and checked for readiness before it is opened officially or expanded to a wider audience.

TOR cannot change, but implementation details can

Patchara confirmed that the project is now in the contract-management stage, meaning the TOR cannot be changed because it is attached to the signed contract.

However, he said implementation details can still be added or refined within each delivery phase to ensure the project delivers clear results.

He said the project would follow the minister’s policy that the state should pay according to actual usage, or “use as much as needed, pay only for actual use”.

Feedback from the public consultation forum will be used to support contract-management details and help adjust project implementation.

Patchara insisted the forum was not being held merely to complete a formal procedure. He said it followed the policy of the minister, deputy minister and government to listen to proposals from all sides and improve the project so it can deliver maximum public benefit.

Contractor explains Pro-level AI and Learn to Earn model

Phakhwan Wongphontawee, director of Human Intelligent Co Ltd, one of the companies in TH Joint Venture, said the contractor had not received any payment from the project.

TH Joint Venture is a partnership between Turnkey Communication Services Plc and Human Intelligent Co Ltd.

Phakhwan said the project is not a free giveaway of AI tools without conditions. Instead, it is intended to let people learn from real use of Pro or Premium-level generative AI tools, which differ clearly from free versions.

She said while many people can already use free AI tools, those tools may not be sufficient for a national project aimed at raising AI capability on a broad scale. Free versions may be limited for deep use, data analysis, continuous work or tasks requiring several AI models together.

She said systematic AI learning requires tools that people can use in practice, not only theory-based courses.

The TH-AI Passport system prepared by the contractor will integrate AI from 14 providers and more than 30 models. In some explanations, the number was given as 31 models.

The models will cover a wide range of functions, including conversation, deep analysis, deep research, image generation, video generation and music creation.

Users will access the tools under learning and point-collection conditions.

Phakhwan said the system will operate under a “Learn to Earn” concept, requiring participants to learn through a Learning Management System in order to earn usage rights.

The course structure is based on 96 original courses under the UNESCO Framework, with additional courses from AI model owners and technology partners. The total now stands at 130 courses.

The courses range from basic AI literacy, safe AI use and prompt writing to report summarising, presentation preparation, data analysis, business use, marketing, sales, customer service and productivity improvement.

Users who complete courses will receive certificates that can be used to support job applications or career development.

Phakhwan said the courses were not generic material that could be found on video platforms. They were designed by specialists and linked to real use of AI tools.

Two user levels and monthly learning conditions

The platform will divide users into two levels: beginners and creators.

New users will receive an initial 100 points to try AI tools on the system. However, if they do not continue learning or collect more points in the following month, their status will return to beginner level and they will not be able to keep accessing all models continuously.

Users must therefore continue learning and collecting points throughout the 12-month project period to maintain broader access.

Phakhwan also defended the project’s cost by comparing it with private platforms that combine several AI models in one system. She said similar services in the market cost around 259 baht per person per month, while the state would pay about 27 baht per person per month under TH-AI Passport.

She questioned whether that could be considered expensive when measured against the number of models, courses, learning systems and usage quota prepared for the project.

Chaolvalit Rattanakornkrisri, deputy managing director for enterprise solutions at Microsoft (Thailand), said the AI platform proposed by the ministry was similar to private services already available in the market.

He said people who wanted to use a similar service on their own might have to pay around 259 baht per month, while under this project the state would cover the cost at about 27 baht per person per month.

The project remains under scrutiny as the ministry prepares to move towards implementation, while the House committees are set to examine the TOR, budget, contract management and procurement process next week.

BangkokBizNews , Posttoday