Thailand’s Electronic Transactions Development Agency, or ETDA, has begun a collaborative process to develop an AI governance and ethics framework for the education sector. The work is aimed at helping schools use AI safely, transparently and fairly while dealing with data protection, algorithmic bias and unequal access to technology.
This is not a new rule that schools or suppliers must comply with today. It is a prototype framework in development. That distinction matters. The practical signal is that AI use in education is moving beyond classroom experiments and toward shared expectations for how tools are chosen, used and overseen.
What changed
According to OpenGov Asia’s report on an ETDA focus group, representatives from Thailand’s Office of the Basic Education Commission, the Institute for the Promotion of Teaching Science and Technology, UNICEF Thailand and Silpakorn University discussed governance standards for education. More than 100 people took part, including school administrators, supervisors, teachers and parents.
The discussions covered AI in classroom support, school management, supervision and student assessment. They also focused on personal data, transparency, bias and the digital divide, especially where children and young people may be affected.
The recommendations are intended to feed into a prototype AI Governance Framework for Education. ETDA says the eventual guidance is meant to be practical for schools, teachers, administrators, developers and agencies that select, build or deploy AI systems.
Why this is an operating question, not just a policy story
Education AI creates decisions that cannot be left to an individual teacher or product team. Who can see student data? When should an automated recommendation be checked by a person? What happens when a tool works better for one group of learners than another? Who is responsible when a system gets something important wrong?
Thailand’s process puts those questions on the table before a finished framework exists. For school leaders, edtech suppliers and organisations running training programmes across Southeast Asia, that is the useful takeaway. Governance needs to be part of the adoption plan, not a document written after a tool is already embedded in teaching or assessment.
ETDA’s wider AI 2026 programme also shows the direction of travel. The agency says it is pairing governance tools with training for government, education, small businesses and communities. Its education work includes AI and digital-ethics training for educators alongside a planned education governance guide.
What operators should do now
Do not wait for a final framework to ask basic governance questions. A small education provider or vendor can begin with a short review of each AI use case:
What learner, teacher or staff data enters the system?
Is the tool advising a person, or making a decision that materially affects a learner?
Can users understand the tool’s role and challenge an incorrect result?
Does the service work fairly for different languages, devices and levels of connectivity?
Who owns the final decision and the escalation path when something goes wrong?
This is not a compliance checklist for Thailand. It is a way to avoid building a workflow that later has to be redesigned because privacy, fairness or human oversight were treated as afterthoughts.
What this may mean for the regional market
Thailand’s initiative does not prove that other Southeast Asian markets will adopt the same framework or timetable. It does show a large regional education system treating AI governance as an implementation problem, not only a national strategy topic.
That is worth watching for organisations that sell into schools, universities, training providers or public-sector programmes. Buyers may increasingly ask for plain answers on data handling, human review, accessibility and how a product behaves in real learning settings. Those questions are useful product work whether or not a formal requirement arrives.
What government and institutions are trying to solve
The stated goal is to let schools gain from AI-supported learning and administration without increasing risks to students or widening unequal access. ETDA’s broader programme describes the same balance in practical terms: build tools and skills that make AI safer and more accountable to use.
For public institutions, the hard part is making that workable for people with very different resources. A framework that only works for well-funded schools will not address the access problem it is meant to solve. That is why teachers, parents, administrators and technical partners need to be part of the design, not only consulted after it is finished.
What individuals and small teams can do
If you are introducing AI into teaching, training or assessment, keep a simple record of the purpose, data used, person responsible and review step for each workflow. If you build or sell an education product, make those details easy for a buyer to find rather than burying them in legal language.
The best early test is modest: choose one low-risk use case, define what a human still reviews, and ask users where the tool could disadvantage someone. That produces evidence a school or vendor can use when governance guidance becomes more specific.
What to watch next
Watch for the prototype framework itself, any operational guidance for schools and developers, and signs that the approach moves from workshops into training, procurement or programme requirements. Until then, treat Thailand’s move as an early operating signal, not a completed regulatory regime.
The useful question is no longer whether AI will reach education. It is whether the people deploying it can explain how it protects learners, supports teachers and remains accountable when the system is wrong.
Source links:
Follow The Workflow Lab for practical Southeast Asia AI signals grounded in what operators can do next.