A student admitted through Posting Group 2 (PG2) in Punggol studies under Full Subject-Based Banding, and digital learning is part of the secondary-school experience regardless of the pupil’s admission group. Singapore’s Student Learning Space (SLS) includes AI-enabled tools for feedback and guided learning, while schools set their own assignment, device and academic-integrity expectations. PG2 does not determine whether a student is allowed to learn about artificial intelligence. The practical rule is that technology should help a child understand and practise a subject, not quietly do the thinking or submit an answer on the student’s behalf.

For 2027, MOE has announced that AI literacy is being strengthened across curriculum and self-directed learning resources, that an updated Code for Fun programme will be available to all schools, and that Cyber Wellness lessons include checking generative-AI information and recognising deepfakes. A Punggol family should begin with the child’s actual school instructions, use approved platforms where provided, protect private information and ask the pupil to verify important answers. An AI explanation that sounds confident can still be wrong; the student’s own reasoning, teacher feedback and official syllabus remain essential.

• PG2 definition: Posting Group 2 is an admission category; G1, G2 and G3 indicate individual subject levels.
• School tools: MOE’s SLS provides digital lessons and certain AI-enabled learning features, subject to teacher setup and availability.
• SLS Learning Assistant: A teacher-enabled conversational tool that guides through questions; its generated responses can be inaccurate.
• SLS Adaptive Learning System: Provides differentiated pathways and feedback, including for lower-secondary Mathematics topics where available.
• 2027 change: MOE’s updated Code for Fun programme will be available to all schools, not necessarily as an identical lesson for every pupil.
• Verification: Check AI claims against textbooks, teachers, official sources, measurements or valid mathematical working.
• Academic integrity: Follow school rules on AI use, disclose assistance where required and submit work that represents actual learning.
• Cyber safety: Do not upload classmates’ private information, share deepfakes, spread rumours or click unknown school-login links.
• Local school operations: Timetables and HBL arrangements differ; Punggol Secondary’s 2026 odd/even timetable is an example, not a universal 2027 schedule.
• Parent action: Establish one safe study workflow: attempt independently, ask for a hint, verify, explain and correct.

Definition lock: AI literacy is not a new academic stream

PG2 helps a pupil enter secondary school. Once admitted, the child studies subjects at the levels actually offered and joins the broader school community. Artificial intelligence, digital communication and information literacy are not reserved for PG3 pupils. A child at G2 Mathematics can learn to ask precise questions, examine computer-generated explanations and decide whether evidence supports a claim. These are reasoning skills that matter across all educational pathways.

AI literacy means understanding enough about AI systems to use them responsibly and critically. It does not mean a thirteen-year-old must become a machine-learning engineer. A pupil can begin by recognising that a generative model predicts and produces plausible responses, may make errors, and does not automatically know whether a school answer satisfies the current rubric. Learning how to verify a result is part of understanding the technology.

The school remains responsible for defining how its assignments should be completed. A parent should not assume that because a chatbot can explain algebra, it may write every submitted answer. Some tasks are designed to assess the pupil’s original thinking; others may permit specified digital tools. The rules depend on the task and school guidance. Sound learning requires a distinction between using a tool to practise and using it to conceal missing understanding.

This article is the AI and cyber-wellness owner in eduKatePunggol’s PG2 series. Its purpose is practical: how to use digital tools without losing the student’s capacity to think, communicate and stay safe.

What MOE actually announced for 2027

In its March 2026 Committee of Supply announcements, MOE said AI literacy would be strengthened through the curriculum, co-curriculum and self-directed learning resources. It also announced that Code for Fun would be updated to build students’ AI-related skills and made available to all schools in 2027. The same announcement said Cyber Wellness lessons had been updated to cover validating information generated by AI and recognising deepfakes.

These are system-level announcements. They should not be repackaged into claims such as “every PG2 pupil will receive the same weekly AI lesson” or “every Punggol school has already installed the same chatbot.” Schools select and implement activities within actual arrangements. Families should obtain current programmes, dates and access instructions from the individual school. An opportunity available to schools is not proof of identical student participation.

The broader educational message is that AI should be studied with both capability and judgement. A pupil needs to know what technology can do, what it cannot reliably do and how its use affects other people. Coding, digital making and evaluating AI outputs can be opportunities for children with different academic and practical strengths.

For a parent, this means looking for substance rather than a fashionable headline. Does the activity teach verification? Can the student explain what the AI did? Does the pupil understand risks involving privacy and bias? Those outcomes are more meaningful than simply reporting that a teenager has used a new tool.

The Singapore Student Learning Space is the proper starting point

The Singapore Student Learning Space is an MOE learning platform used to support teaching and independent work. Its official information pages describe several AI-enabled features, but access to a particular activity depends on school and teacher setup. Parents should not assume that every named feature is present in their child’s current homework simply because it appears on an MOE website.

The Adaptive Learning System offers personalised learning pathways and feedback for supported topics, including upper-primary and lower-secondary Mathematics and upper-secondary Geography. For a PG2 pupil struggling with a lower-secondary Mathematics concept, a teacher-assigned or available adaptive activity may help diagnose which steps have been understood. This is different from a random worksheet that offers the same questions regardless of readiness.

SLS also describes a Learning Assistant, feedback assistants and other tools for particular educational jobs. The Learning Assistant can guide a student through iterative questioning. Feedback features may point to errors or suggest improvements. Teachers can use some tools to design learning activities and monitor understanding. The platform’s official information states that these features are governed by MOE’s AI arrangements and reviewed for fit and safety.

No automated tool replaces the role of the teacher. The student still needs to interpret the task, think independently and ask for help when a response does not make sense. SLS is most valuable when it extends good teaching rather than becomes another place to copy answers.

The Learning Assistant: a dialogue is not a perfect answer key

MOE’s SLS information identifies the Learning Assistant as a student-facing dialogic agent. It is designed to help students learn through questions and responses. The official student guide updated in October 2026 warns that generative AI outputs can be inaccurate and encourages students to verify the content with other sources. It also instructs pupils to complete the relevant basic AI modules before using the feature.

A good interaction begins with the child’s own attempt. Suppose a G2 student is unsure why 3(x + 2) expands to 3x + 6. Instead of asking for a finished worksheet, the pupil can request a hint about what the multiplier applies to. The student should then draw or write three identical groups of x + 2, count the terms and explain why each part is multiplied. The tool may prompt the reasoning, but the learning occurs when the child can complete the next example unaided.

Teachers may configure the feature with learning materials and interaction limits. A parent should not treat the tool as an unrestricted private tutor guaranteed to know every local school’s exact worksheet. If a generated explanation conflicts with the textbook or teacher’s method, the student should check the authoritative material and ask the teacher for clarification.

This is a useful lesson beyond examinations. A conversational system can sound certain even when its conclusion is not justified. Learning to test a claim is a foundation of digital citizenship.

The five-step rule for using AI without outsourcing thought

First, attempt the problem independently. Even a partial attempt gives the learner a starting point and reveals the difficulty. Second, ask for the smallest useful help, such as a clue about the relevant principle or a question that identifies the missing step. Third, compare the generated suggestion with a trusted source and the actual task. Fourth, close the tool and explain the solution in the child’s own words. Fifth, solve a changed example without help.

This workflow is deliberately demanding. Copying a perfect-looking answer may create the impression of success without altering the student’s understanding. A student who cannot explain the output is not yet ready to claim mastery. A thoughtful tutor or parent should value the independent second attempt more than the polished first response.

For English, the tool might help the child distinguish a claim from evidence. The student still chooses the quotation and writes the explanation. For Mathematics, it might point out an algebraic misconception, but the learner must verify each step. For Science, it can help organise a comparison, but factual mechanisms should be checked against class materials.

If the child repeatedly needs the complete answer before proceeding, reduce the task to a smaller concept and seek human guidance. More sophisticated prompts are not a substitute for teaching an essential foundation.

AI in G2 English: useful editing versus hidden authorship

English students can use digital tools to practise grammar, compare sentence structures and generate questions about a passage, subject to school rules. A pupil might write a paragraph and ask for feedback on whether the topic sentence matches the supporting evidence. The most useful feedback identifies a specific issue and encourages the student to revise it independently.

The risk arises when the pupil asks AI to produce the whole composition and submits it as original work. The result may look fluent, but the student has not practised planning ideas, choosing vocabulary or controlling sentences. A grade attached to the work would no longer accurately represent the child’s learning. There may also be an academic-integrity breach under the school’s instructions.

Try this exercise at home. Ask the teenager to explain a passage aloud, write a short response without assistance, then use permitted feedback to identify one weakness. The child revises the work and explains the change. When successful, the new paragraph is still recognisably the pupil’s writing rather than a generic machine-produced essay.

Parents should avoid rewarding only polish. An imperfect but genuine attempt may reveal the exact teaching needed. If adults praise only the most elegant product, a child may learn to hide uncertainty instead of developing language ability.

AI in Mathematics: check the working, not the confidence

Mathematics is particularly vulnerable to persuasive but invalid automated steps. A system may present a neat calculation or a plausible answer without correctly following the problem’s constraints. The pupil should therefore check operations, signs, units and assumptions. At G2, the habit of showing and verifying working is more valuable than an unexplained final number.

Consider the expression 2(x + 4). The correct expansion is 2x + 8. A child who writes 2x + 4 has not distributed the multiplier to both terms. A good hint leads the learner to represent two identical groups and compare the result. An unhelpful automated answer simply supplies 2x + 8 without making the reason visible. The student might copy it and repeat the error tomorrow.

When a chatbot provides an unfamiliar method, ask the pupil to verify it using substitution or another trusted approach. If both methods are valid, the comparison can develop flexibility. If the answer conflicts with class teaching, determine where the disagreement begins. One worked example from the teacher or textbook may reveal an omitted condition.

Do not assume every AI mathematics assistant can read handwritten notation perfectly or interpret an image without error. Transcription problems can change coefficients, signs and exponents. The pupil should read the input and output carefully before treating the result as evidence.

AI in Science: facts, explanations and invented citations

A science answer can sound impressive while containing a factual error or an unsupported causal explanation. Students should learn to identify the claim, the relevant evidence and the mechanism that connects them. An AI-generated paragraph can help suggest questions, but it is not a replacement for the prescribed syllabus, classroom experiment or verified reference.

Suppose a student asks why a plant grows differently under two conditions. A useful response should distinguish the variable changed, the observation made and a plausible mechanism supported by the experiment. If the tool asserts that the result was definitely caused by one factor without accounting for controls, the pupil should challenge that inference. Scientific thinking includes recognising what the experiment cannot prove.

Source checking is equally important. A chatbot may provide a reference that looks scholarly but is inaccurate, incomplete or does not support the quoted claim. Teach the student to open a cited page and verify that it actually says what the answer alleges. A bibliography is not proof merely because it looks formal.

Encourage the pupil to write the explanation in their own words after checking the evidence. If they can describe the mechanism without reopening the tool, the digital interaction may have supported real learning.

Humanities and Social Studies: learning to challenge a plausible narrative

History and Social Studies require careful interpretation of evidence, perspective and context. A fluent AI answer may summarise a historical issue from one point of view without indicating uncertainty. Students should learn to ask who produced a source, what claims it supports, what it leaves out and which independent evidence would help verify the conclusion.

Imagine an AI system giving a confident account of a controversial event and providing three unsourced assertions. The child should not assume the tone indicates accuracy. Ask what dates, documents, institutional sources or credible research would allow those statements to be checked. The same habit applies to current events and local rumours about Punggol schools or government projects.

For source-based work, a student can use a permitted tool to practise identifying a claim and counterargument. But the pupil must still interpret the actual material supplied in the assessment and explain how the evidence supports a conclusion. A ready-made generic paragraph can be irrelevant even when every sentence is grammatically correct.

The educational target is independent judgement: not cynicism about everything, and not unquestioning trust in a machine. Good reasoning requires proportionate confidence based on evidence.

Code for Fun and the difference between using and understanding AI

MOE’s 2026 announcements describe an updated Code for Fun programme available to all schools from 2027 to strengthen AI skills. IMDA’s programme information explains the wider emphasis on computational thinking, digital making and technologies such as AI, with responsible use as part of the learning. This is relevant to a Punggol pupil who enjoys building, coding, solving problems or exploring how technology works.

Parents should ask what activity the child actually completed. Did the pupil design a simple sequence of instructions, test a program, recognise an error or explore a real-world application? These are more meaningful than simply reporting that an AI tool generated a picture or a chatbot answered questions. The process is where computational thinking develops.

A PG2 pupil can contribute through curiosity, careful testing, design choices or the ability to explain a solution to peers. Coding and AI literacy do not belong exclusively to pupils with a high PSLE score. At the same time, a project-based activity should not be used to claim a teenager has mastered advanced computer science. Describe the actual skill learned accurately.

Punggol Digital District and the nearby Singapore Institute of Technology provide a local reminder that digital technologies are connected to real working environments, but their presence does not create an automatic school programme or guaranteed university pathway. Use the local context for curiosity, not a fictitious promise.

Digital citizenship begins with privacy

Students should understand that school work, private conversations and personal information are not all appropriate material for a public AI chatbot. Do not paste classmates’ full names, telephone numbers, photographs, identity details, medical information or sensitive school messages into an unapproved service. Even when a task sounds harmless, unnecessary disclosure can harm another person.

Teach a practical anonymisation habit. If the child wants help organising a study schedule, the tool does not need the names of all classmates or a screenshot showing their contact details. If the student is asking about an English paragraph, private information in the draft can be replaced with neutral examples. If an assignment includes restricted data or a real pupil’s story, follow the teacher’s instructions rather than uploading it to a third-party tool.

Passwords and account recovery information should not be shared with friends or entered into a link received through an unexpected message. Students should use official school login routes and ask the school when uncertain. Parents should also avoid publicly posting screenshots of grades or school portals that reveal personal details.

Privacy is not a rule that only adults understand. A thirteen-year-old can learn to ask: “Does this tool genuinely need the information I am about to provide?” That question is useful in school, at work and throughout life.

Deepfakes and the problem of believable images

A deepfake is manipulated or generated media that can falsely represent a person or event. It may involve images, audio or video. The central danger is not merely that a picture looks strange; it is that it may look entirely plausible while misrepresenting reality. MOE’s updated Cyber Wellness curriculum includes recognising deepfakes and checking AI-generated information, reflecting how important these skills have become.

When a teenager sees a dramatic clip about a classmate, teacher or public event, the correct first response is not to share it urgently. Check the original source, look for corroboration, consider whether the account is trustworthy and ask whether the context may have been changed. Visual appearance alone is not reliable proof. Even authentic footage may be edited or framed to give a misleading impression.

Students should also understand consent. Creating a fake embarrassing image of a schoolmate is not a harmless test of technology. Sharing such content can cause serious harm, even when the creator claims it was a joke. School rules and legal consequences may apply depending on the conduct. The safe choice is to avoid creating or distributing non-consensual harmful material and to report an incident through appropriate channels.

Parents can practise media verification with non-personal, harmless examples, asking the pupil to identify what is known, unknown and independently verifiable. The skill is judgement, not merely detecting odd pixels.

Cyberbullying, group chats and AI-generated harassment

A school-related group chat can become harmful when false rumours, impersonation, threatening messages or humiliating images are spread. AI tools can make the problem worse by producing convincing fake screenshots or manipulated media. The student needs to know that digital actions have real social consequences and that reposting harmful content expands its audience.

If a pupil encounters such material, they should avoid forwarding it, seek help from a trusted adult and use the school’s appropriate reporting route. Preserve necessary evidence privately where safe and permitted, then block or report the harmful account through the platform. For urgent threats to safety, contact school staff or relevant authorities directly rather than relying only on an online reporting form.

Parents should not respond by publishing other children’s identities or alleged wrongdoing on social media. Doing so may expose pupils, spread unverified claims and make an investigation more difficult. A careful account of dates, incidents and impact is more useful for the school than a viral accusation.

Digital literacy includes learning when not to participate. A teenager who chooses not to share an unverified image has exercised judgement, even if classmates say that everyone else has already seen it.

The school’s device rules are more important than a generic recommendation

Different secondary schools may specify learning devices, management software, platform access and acceptable-use rules. Punggol Secondary School’s published 2026 Semester 2 timetable, for example, described odd and even weeks with Home-Based Learning once a fortnight. That is a dated local example, not a promise that all Punggol schools or all 2027 cohorts use the same arrangement.

When the pupil receives current school instructions, check whether the required device has been specified, how learning platforms are accessed and what activities are allowed. Families should not purchase extra hardware simply because a website claims that every student needs a particular specification. Schools may also provide information about financial assistance or approved device arrangements; ask through the official channel.

An HBL day calls for its own routine. The pupil needs to know where assignments appear, when to complete them, how to seek help and what to do if connectivity fails. A device that is fully charged but used for hours of unrelated content is not evidence that learning occurred. Conversely, a student who completes a shorter task attentively may learn more than one who remains online all day.

Parents can help by designing a quiet study space, encouraging breaks and checking that the child can explain what they have learned. Monitoring should be proportionate and build responsibility rather than total surveillance.

AI study plans should not turn every evening into a screen session

It is possible to overuse even an educational tool. A student may spend an hour asking a chatbot to revise the same study schedule, produce colourful summaries or reassure them that an assignment is excellent. The activity looks academic, but it may contain very little independent retrieval, writing or problem-solving. Parents should judge learning by what the child can do when the screen is closed.

Try a simple test. After using AI assistance for a concept, ask the pupil to explain it on paper and solve a changed question without prompts. If the child cannot, return to a smaller principle or school teaching material. This is not a punishment. It is a way to determine whether the tool supported understanding or merely produced attractive output.

Study time should include offline reading, written working, discussion and rest. Some assignments legitimately require online research or a digital platform, but constant connectivity is not the same as concentration. Notification rules and a predictable homework routine can reduce interruptions without demonising technology.

An AI-assisted learning plan should also respect the child’s age and daily commitments. CCA, transport, friendships and sleep matter. If a new tool adds complexity rather than reduces it, simplify the workflow.

How to teach prompt quality without turning prompts into tricks

A useful question to AI identifies the topic, the learner’s attempt and the exact uncertainty. “Explain Mathematics” is vague. “I can solve 2x = 10 but do not understand why 2(x + 3) must multiply both terms; ask me one question at a time” provides a more useful learning context. The second request also encourages the student to remain engaged rather than receive a complete answer immediately.

However, prompt quality is not an academic shortcut. A beautifully phrased request cannot make an incorrect generated answer true. The student still needs to verify concepts and reproduce the reasoning. Learning how to ask a clear question is valuable partly because it transfers to conversations with teachers and tutors.

Encourage pupils to include their attempted working rather than hiding it. “My answer is 2x + 3; which step is wrong?” exposes the misconception. A hint can then focus on the distribution rule. For an English response, “I wrote this inference but cannot find evidence in the passage” identifies an interpretable problem. The pupil still has to choose and explain the evidence.

Avoid teaching students to manipulate tools into appearing to have done original work. The ethical aim is clearer understanding and better questions, not undetectable outsourcing.

Three fictional Punggol learning situations

The algebra shortcut. A pupil studying G2 Mathematics uploads a question and receives a correct answer. The child copies the result but cannot explain why the brackets were expanded. A parent asks for a second example with changed numbers, and the pupil discovers the same misconception. The family uses a diagram and teacher guidance to explain distribution, then checks whether the student can solve the next problem independently. The tool becomes a source of clues rather than a substitute for learning.

The convincing invented science explanation. A teenager asks a chatbot why a particular experiment produced an unexpected result. The tool offers a plausible mechanism but cites no usable evidence. The student compares the answer with the textbook, checks which variables were actually controlled and writes a conclusion limited to the observations. The important success is not the length of the AI response; it is the pupil’s willingness to test the claim.

The suspicious group-chat image. An image appears to show a classmate making an offensive statement. Several friends want to share it. The pupil notices that the origin is unclear and does not forward it. They tell a trusted adult if the content is harmful and let the appropriate school processes address the matter. The student’s critical thinking has prevented an unverified claim from spreading further.

These are hypothetical teaching examples. They do not describe actual incidents at Punggol schools and should not be interpreted as local allegations.

A family agreement for responsible AI use

A short agreement can help everyone know what is expected. The pupil agrees to attempt school work independently first, use only permitted assistance, check important AI outputs and disclose tool use where required. The parent agrees not to demand polished work that could encourage concealment and to seek teacher guidance when assignment rules are unclear. Both agree not to upload unnecessary personal information or forward harmful synthetic media.

Discuss what happens after an error. If the child used AI inappropriately, the aim is to address the conduct and repair the learning. Ask which part of the assignment they cannot yet complete, explain the academic-integrity problem and follow school procedures if relevant. A constructive response makes future honesty more likely.

Review the agreement as school requirements change. A Secondary 1 pupil may need more supervision than an older student completing complex research. The purpose of the rules is not to keep a teenager dependent forever, but to develop judgement that works when no parent is looking over the screen.

Keep the agreement simple enough that the family actually follows it. A dozen meaningful habits are more useful than pages of restrictions nobody reads.

Where three-student teaching can complement digital feedback

At eduKate, the small-group tutorial idea is to observe how each learner reasons. Digital tools can provide extra practice or hints, but a human tutor can watch the pupil’s hesitation, ask why they chose a step and notice when an explanation is being repeated without understanding. In a three-student group, pupils can compare methods and explain corrections to one another while receiving focused feedback.

This matters particularly when AI supplies a fluent answer that the child cannot evaluate. A tutor can request the student’s own attempt, identify where the mental model broke and design a different problem to test learning. For English, the tutor can ask why one quotation supports an inference. For Mathematics, they can inspect handwritten algebra. For Science, they can check whether the pupil understands an experimental limitation.

Tuition is not required simply because AI exists or because the student is PG2. School teaching and independent practice may be enough. Additional support is justified when it addresses a real, persistent learning need, not when it creates another channel of worksheets or automated answers.

The goal should be a student who becomes less dependent on both humans and machines for routine thinking and more capable of asking precise questions when genuinely stuck.

Frequently asked questions about PG2 and digital learning

Are AI learning opportunities limited to PG3 pupils?

No. PG2 is a Secondary 1 posting group, not a digital-literacy category. Access to a particular learning activity depends on the school, teacher and platform arrangements, not a universal assumption about the pupil’s admission label.

Is SLS the same as a public chatbot?

No. SLS is an MOE learning platform with educational features governed by school and MOE arrangements. Some SLS features use AI; access and setup vary. Public chatbot services have their own terms, privacy practices and school-use restrictions.

Can SLS Learning Assistant give wrong answers?

Yes. MOE’s own guide says generative AI responses may be inaccurate. Pupils should verify important information and use teacher-approved materials and explanations.

Does MOE require every student to attend the same Code for Fun course in 2027?

MOE announced that the updated programme would be available to all schools. That does not establish an identical schedule or activity for every individual pupil. Check the school’s actual programme information.

Can a pupil submit a fully AI-written essay?

Follow the school’s specific academic-integrity rules. Submitting machine-written work as independent original work can undermine assessment and violate instructions. AI is better used for permitted feedback or practice that leaves the student responsible for the final thinking.

Is an AI-provided citation always reliable?

No. Open the cited page and check that it exists and supports the claim. Treat a plausible-looking citation as an item to verify, not proof.

Can students upload photos of classmates to an AI tool?

Not without a legitimate, permitted reason and appropriate consent or school authorisation. Avoid unnecessary sharing of identifiable or sensitive personal information.

What should a student do after receiving a deepfake?

Do not spread it. Check the source, preserve necessary evidence privately if appropriate and report harmful or threatening content to a trusted adult and the relevant school or platform channel.

Do all Punggol schools use the same HBL timetable?

No. Punggol Secondary published an odd/even timetable for 2026, but another school or later cohort may have different arrangements. Follow the student’s current timetable and school notices.

Does using AI help a child move from G2 to G3 automatically?

No. Subject-level progression depends on school criteria and demonstrated readiness. AI may support useful practice, but cannot guarantee a change in academic level.

How should a parent know whether AI helped?

Ask the pupil to explain the idea and solve a new problem without assistance. Look for greater accuracy, independent reasoning and the ability to identify and correct mistakes.

What is the best first family rule?

Attempt independently, request only necessary help, verify the answer and explain it without the tool. Combine that learning rule with strict protection of personal information and compliance with school instructions.

Your next step in Punggol

Open the student’s actual school learning instructions and identify which digital platforms and AI features are permitted. Agree on a short study process, practise checking a generated claim and set clear privacy boundaries. Ask the pupil to demonstrate understanding on paper after using any tool. Digital competence is not about collecting the most applications; it is about using technology thoughtfully without surrendering judgement.

This is the AI and cyber-wellness page in the PG2 series. For the first-school-term routine, read the dedicated 100-day guide; for G2/G3 subject learning, read the individual subject-level guide; for social or bullying concerns, read the PG2 wellbeing guide. Each has a different purpose so that parents can find relevant information without repeated generic essays.

Official sources and local reading

• MOE: Committee of Supply 2026 AI literacy and Cyber Wellness announcements
• Singapore Student Learning Space: AI-enabled features
• SLS: Guidance on generative AI, updated September 2026
• SLS: Student guide to Learning Assistant, updated October 2026
• SLS: How AI features are governed
• IMDA: Code for Fun programme
• Punggol Secondary School: 2026 timetable and HBL arrangements
• What is PG2: First 100 days in Punggol secondary school
• What is PG2: Mixed form classes and wellbeing