Student Learning Space: How Singapore is integrating AI tools into its national education platform

The problem

Teachers in Singapore's national school system carry a broad set of responsibilities alongside classroom instruction. They plan lessons, design assessments, mark student work, write feedback, track each student's progress, and produce detailed testimonials. Much of this work is repetitive and time-consuming and requires care because it directly affects how students learn and how their progress is recorded.

The shift to home-based learning during the COVID-19 pandemic in 2020 accelerated the need for more capable online learning tools. Schools needed a platform that could support students learning remotely while also helping teachers manage the increased workload of planning and delivering lessons across digital and classroom settings.

At the same time, advances in artificial intelligence were opening up ways to automate some of the more routine parts of teaching, such as generating first drafts of lesson plans, providing preliminary feedback on student answers, and identifying patterns in how students were responding to material. The question for Singapore's Ministry of Education (MOE) was how to bring these capabilities into schools in a way that supported teachers rather than bypassed them.

The solution

Singapore's MOE, working with GovTech Singapore, the government's technology agency, added a set of AI-powered features to the Student Learning Space (SLS), the national online learning platform it launched in 2018. SLS is available to all teachers and students in the national school system and offers curriculum-aligned resources for the main subjects across primary, secondary and pre-university levels, updated over time with input from teachers and students.

Several AI-enabled tools have been added, each addressing a different part of the teaching process.

The Adaptive Learning System (ALS) uses machine learning, software that refines its recommendations over time from the data it processes, to build individualised learning paths for students. It is currently available for Mathematics (Upper Primary and Lower Secondary) and Geography (Upper Secondary). ALS adapts what it shows, the questions it sets, and the comments it returns based on each student's progress and grasp of the topic. When a student answers incorrectly, it can offer a hint or explanation before they attempt the question again. Students can also use ALS on their own to revise topics or prepare for tests, and teachers see a dashboard of each student's progress and areas of difficulty that helps them decide where to focus.

The Authoring Copilot (ACP) helps teachers turn lesson ideas into structured digital lessons. It runs on a large language model, an AI system that can read and generate text. A teacher describes what a lesson should cover, and the tool proposes a lesson structure with components such as multiple-choice and open-ended questions, polls and discussion activities. Teachers can upload source material, such as a textbook chapter, for the tool to build activities from, then review and edit the output before it is used. For now, the tool handles text; MOE has said planned additions include pictures, scanned documents and video transcripts.

The Short Answer Feedback Assistant (SAFA) drafts comments on students' written answers. The teacher first sets out what a strong answer should contain and how marks are awarded; the tool then produces a draft comment and a suggested mark for each response, which the teacher reviews, adjusts if needed and approves before the student sees it. It covers most subjects and levels, but not maths questions that require step-by-step working to be marked.

The Data Assistant (DAT) analyses how a whole class has responded to questions, discussion prompts and interactive exercises, helping teachers see where a class has gone wrong and what is recurring across answers. Teachers can group students by performance and add targeted comments for each group, and DAT connects with SAFA so that individual feedback feeds into a class-level view.

Outside SLS, GovTech and MOE also built the Appraiser Testimonial Generator, which helps teachers draft the personalised testimonials each student needs for university and scholarship applications. Writing these by hand takes teachers many hours. Appraiser uses a large language model, OpenAI's ChatGPT, to produce a first draft in minutes, which the teacher then reviews and personalises.

Across tools, a teacher must review AI-generated feedback before a student sees it, and can amend it or provide their own.

SLS also hosts digital literacy modules, part of Singapore's National Digital Literacy Programme, that build students' technological and new-media skills and cyber-wellness.

These efforts fall under MOE's EdTech Masterplan 2030, launched in September 2023, which sets out how technology should help schools meet a wider range of learning needs and prepare students for a digital economy.

The impact

1. The Appraiser tool has generated tens of thousands of testimonials.

As of early 2025, more than 4,000 teachers had used Appraiser to generate over 40,000 testimonials. Writing individualised testimonials for university and scholarship applications takes teachers many hours. The tool produces a structured first draft in minutes, which the teacher then reviews and personalises.

2. The tools are available to every teacher and student in the national school system.

SLS is not a pilot in selected schools. It reaches every teacher and student across the school system at every level up to pre-university, and the AI features are available to any teacher using it.

3. The Adaptive Learning System provides individualised learning paths in two subjects.

ALS currently covers Upper Primary and Lower Secondary Mathematics and Upper Secondary Geography, with plans to expand. Students receive content and practice questions that adjust to their progress: if they are struggling, the system offers more support before moving on; if they have understood, it advances them. Teachers receive a dashboard showing where each student stands, and the system is built to scale as new subjects are added.

4. The platform received independent international recognition.

SLS was awarded the ISTE Seal by the International Society for Technology in Education following a rigorous evaluation. MOE describes the Seal as recognising usability, pedagogy, and alignment with the ISTE Standards. It states that the evaluation confirmed the quality of the platform's design and teaching approach, as well as its support for inclusive learning.

Lessons learned

• The approach depended on a national platform that was already universal. SLS reached every teacher and student because it had been built up across the school system since 2018, and the AI features were added only once it was in place. A new tool could therefore reach every school as soon as it was enabled. Establishing the platform took years, and the AI features came later and built on top of it.

• Human review is built in as a requirement, which shapes how much time the tools save. A teacher checks every AI output before a student sees it, which addresses the concern that AI-generated content might reach learners unchecked. It also means the tools speed up the drafting of feedback, lessons and testimonials without taking the task away from the teacher, who still reads, corrects and approves each result. The saving falls at the drafting stage, and the review stage remains.

• The published risks were paired with an actual test in at least one case. GovTech has set out the risks it sees in using AI in education, including inaccurate outputs, bias, overreliance, and student data privacy concerns. For the testimonial tool, GovTech and Resaro went further and ran a fairness test, building a synthetic dataset and checking for bias by student gender and ethnicity. Here, one of the named risks was examined with a concrete test.

• The design uses a separate tool for each task. The tasks of drafting a lesson, marking a short answer and writing a testimonial each have their own tool, introduced individually as each was built and tested. This lets teachers take up only the tools relevant to their work and review each one against a task they already know well.

This case study was written with assistance from artificial intelligence.