A student encounters a dramatic message claiming that school examinations have been cancelled. Another sees a video of a public figure apparently announcing a policy. A third asks an AI assistant for a fact and receives a fluent answer with an unfamiliar citation. The difficult task is no longer finding words on a screen; it is establishing whether the information deserves belief, use or circulation. Media literacy is the set of habits that makes such judgments more reliable.
Under Full Subject-Based Banding, information verification crosses G1, G2 and G3 English, Humanities, Science, Mathematics, Character and Citizenship Education and digital learning. A student at an introductory level can check who sent a message and where it originated; another learner can investigate whether numerical evidence supports a causal claim; an advanced task may require comparing the limitations and incentives of several sources. The skill is not an extra SEC paper, and it should never be used to rank young people by their subject labels.
This guide develops the National Library Board’s S.U.R.E. framework—Source, Understand, Research and Evaluate—into an everyday school procedure. It also addresses generative AI, manipulated images, emotional headlines, social media algorithms and ethical sharing. MOE’s 2026 announcements identify the validation of AI-generated information and the recognition of deepfakes as part of updated Cyber Wellness learning. The purpose is neither blanket distrust nor naive acceptance. It is a disciplined willingness to ask what the evidence supports and what remains unknown.
The sixty-second S.U.R.E. answer
Check the source, understand the precise claim, research beyond the original post and evaluate the evidence. Before acting or sharing, check dates, original context, independent confirmation and potential harm. AI-generated text, images and deepfakes require the same evidence-first discipline. A convincing presentation or viral headline is not proof.
The practical verification checklist
- Who first produced the claim and where?
- What exactly is asserted, and about what time?
- Is the source authentic and appropriate to the claim?
- Can primary or independent evidence confirm it?
- Are quotations, images and charts shown in context?
- Which uncertainty remains after checking?
- Would sharing now create avoidable harm or confusion?
1. Information literacy is a decision skill
Students rarely encounter information as a tidy classroom fact. A claim may arrive in a group chat, a short video, an attractive infographic or an AI-generated paragraph. The learner must decide whether to believe it, use it in homework, act on it or forward it. Each action has a different cost when the information is wrong.
Teach children to name the decision before verifying. Is the message just an interesting idea, or will someone spend money, miss school, change medication or share personal information because of it? Higher-consequence claims deserve stronger checks and guidance from appropriate adults or professionals.
2. S.U.R.E.: Source, Understand, Research, Evaluate
The National Library Board’s S.U.R.E. framework structures information discernment around four actions: identify and inspect the source, understand what the information claims, research beyond the first source, and evaluate the evidence and perspectives. It is a memorable process but becomes useful only when students apply it to real examples.
Give the learner a fictional claim and ask for one action under each letter. The exercise should not end at finding a logo; an official-looking graphic may be copied or altered. The final result should explain what was verified, what was not and whether sharing or acting on the claim is justified.
3. Source is a question, not a decorative credit
A source may be an eyewitness, an official organisation, a researcher, a journalist, a company or an anonymous account. The relevant question is what the source was in a position to know and whether the claim can be traced to original evidence. A confident tone or a large audience does not substitute for that relationship.
Ask who first made the statement, when and through which channel. Then see whether the cited document actually contains the claim. Sometimes an apparently authoritative article merely repeats another unverified post. Trace the evidence rather than counting how many accounts repeated it.
4. A lookalike website can borrow authority
Fraudulent or misleading pages can imitate familiar visual styles, logos and domain names. Students may notice a government crest and assume authenticity without checking the address. A copied design does not prove that an organisation published the page.
Teach learners to navigate to an official site independently and search for the announcement there. Do not trust a login link merely because it appeared in a forwarded message. If the issue involves urgent school arrangements, verify through the school’s established channels or a responsible adult.
5. Understand the exact proposition
A claim such as ‘students are doing better’ is too vague to verify until we know which students, which measure, compared with what and over what period. Emotional headlines often skip these details. Students must learn to translate a general message into a specific proposition that could, in principle, be tested.
Ask the learner to rewrite the claim with its subject, action, time and evidence requirement. If the original message does not provide enough detail, that absence is itself relevant. Precision comes before the search for confirmation.
6. Separate observation, inference and opinion
An observation reports what was recorded, an inference draws a conclusion from evidence and an opinion expresses a judgment or preference. One message can contain all three. Students may mistake a confident interpretation for an observed fact, especially when the speaker presents it in a polished visual format.
Use a fictional example: ‘The queue lasted twelve minutes’ is a measurement; ‘the service has become less efficient’ is an inference requiring comparison; ‘the queue was unacceptable’ is a judgment. Ask what additional evidence each would need. The categories improve both Humanities source analysis and everyday discussion.
7. Publication date matters
A true announcement from an earlier year may be inaccurate as guidance for the current year. School admissions, examination schedules and technology requirements can change. Social media forwards often strip away the original publication date, causing an old item to appear new.
Teach the student to find when the source was published and whether the claim describes an event, a proposal or a current rule. When checking SEC requirements, for instance, the cohort and examination year must be specified. An authentic old notice is not necessarily suitable for a new decision.
8. Event date and article date are distinct
A news article may be published today while describing an incident from years ago, or it may update a long-running story with one new detail. Readers who focus only on the publication timestamp can misunderstand when the event actually happened.
Ask students to construct a short timeline: when the event took place, when evidence became available, when the article was published and whether a correction followed. This habit is valuable in History and in evaluating fast-moving online information.
9. Research beyond the first result
Search engines rank material for relevance and other factors, not simply truth. A high-ranking page can be outdated, copied or misleading. Students should compare independent and authoritative sources relevant to the claim rather than assume that the first result answers the question.
Teach lateral reading: leave the original page, investigate who operates it and look for primary documents or reputable corroboration. Return to the claim with a better understanding of the source. The method is especially useful when a site looks professional but offers little information about its authors or evidence.
10. Repeated copies do not create independent confirmation
Several websites may repeat the same unsupported assertion. Counting them as multiple independent sources gives a false impression of corroboration. A claim copied through a network remains one original claim unless additional evidence has been collected.
Ask the learner to identify whether each report adds a new document, measurement, eyewitness account or expert analysis. If all point back to the same anonymous message, say so. Independent confirmation depends on distinct evidence, not the number of times a sentence appears online.
11. Primary sources still require interpretation
A primary source is close to an event or original record, but it may be incomplete, mistaken or self-interested. A public statement can reliably tell us what an organisation said without independently proving every claim contained in the statement. Students should distinguish authenticity from accuracy.
Teach the two questions separately: Is the document genuinely from the named source? Does its content establish the claim being made? A student’s own photograph may authentically show one moment while revealing little about events before or after it.
12. Photographs show selected frames
A genuine photograph can mislead through cropping, timing or missing context. A crowded image may show a peak moment rather than an average condition. Cropping may conceal a sign or nearby activity that changes interpretation. The image need not be digitally altered to create an inaccurate impression.
Ask what lies outside the frame, when the photograph was taken and whether other images or records support the narrative. Avoid inventing an explanation from visual clues alone. Responsible interpretation often begins by saying what is visible and what remains uncertain.
13. Reverse-image checking can identify reuse
When appropriate tools are available, checking whether an image appeared earlier in another context can reveal that an old photograph is being presented as a new event. This method is useful but not definitive: an image may be new, edited or absent from indexed collections.
Teach the limitation as well as the technique. If no match appears, that does not prove authenticity. If an earlier appearance is found, inspect whether the location and context were accurately described. Verification combines clues rather than relying on one search result.
14. Deepfakes change what seeing means
Synthetic video and audio can imitate voices, faces and situations. A convincing clip is therefore not proof that the depicted statement was genuinely made. Some genuine videos can also appear unusual because of compression or editing, so visual glitches alone are unreliable evidence of fabrication.
Use official confirmation for high-consequence claims. If a clip announces a school rule or demands payment, seek the announcement through a known school or government channel. The goal is a reliable authentication process, not teaching children to make confident judgments from facial details alone.
15. AI answers can be fluent and wrong
A generative AI response may provide a useful explanation while inventing facts, attributing a claim to the wrong source or giving outdated rules. The style of the answer is not a quality guarantee. Students can use AI to generate hypotheses and questions, but important facts need appropriate verification.
Ask the learner to extract specific claims from the answer and check them against primary documents or trusted materials. If the citation is absent, broken or unrelated, do not use it as evidence. The student should be able to state where each important factual assertion came from.
16. AI-generated citations need to be opened
A formal-looking reference is not proof that a real document exists or supports a sentence. Students may see an author, title and year and assume scholarly authority. They should verify that the source exists, that the cited passage is relevant and that it has not been distorted.
Practise with a harmless fictional citation and ask what would need to be checked before including it in an assignment. Teach that an empty citation list is better than fabricated attribution. If a fact cannot be confirmed, the honest response is to qualify or omit it.
17. Algorithms shape what appears in a feed
Social platforms select and arrange content partly to influence engagement and user experience. What a teenager sees often reflects prior interaction and ranking systems, not a neutral sample of public opinion. A feed full of one claim does not establish that most people agree with it.
Ask students to consider what content may be missing from their view and how repeated exposure affects familiarity. Encourage active searches for relevant counterevidence and primary sources when the issue matters. Awareness of personalisation helps readers avoid mistaking their own media environment for the whole world.
18. Emotion can signal the need to slow down
Messages designed to provoke anger, fear or outrage can spread rapidly because people feel compelled to react. Strong emotion does not prove a claim false—real events can be disturbing—but it raises the importance of checking before sharing. Urgency is especially dangerous when a post requests payment or personal details.
Teach a pause routine: identify the factual claim, locate its source and consider what harm sharing an unverified claim could cause. Students should not be scolded simply for feeling an emotion; they should learn how to separate that reaction from the decision to act.
19. Headlines may overstate the evidence
A headline compresses a story and may omit qualifications, sampling limitations or uncertainty. Students who share it without reading further can accidentally circulate a claim stronger than the article itself supports. An image attached to the headline may also be illustrative rather than evidence of the described event.
Ask the learner to write the claim made by the headline and then compare it with the actual article body. Does the article establish causation or only association? Does the claim apply to a small sample? The difference between summary and evidence matters in every subject.
20. Statistics require denominator and source
A number can be accurate yet misleading when its denominator, population or time period is missing. ‘Most respondents agreed’ might describe a voluntary poll among a narrow group. Students should ask who answered, how they were selected and what question they were asked.
Connect this check with Mathematics and Geography. Reconstruct a simple fictional survey table and compare the numerical proportion with the headline. If the table does not support a national generalisation, revise the statement. Quantitative literacy and media literacy reinforce each other.
21. Correlation is not automatically causation
A report may claim that one behaviour causes better results because the two appear together. Other factors may be involved, and a graph alone does not identify the causal mechanism. This is especially relevant to claims about learning methods, health outcomes or consumer products.
Ask what alternative explanation would also fit the observations and what kind of evidence could discriminate between them. Do not dismiss a plausible claim reflexively; instead calibrate the conclusion to the available evidence. A cautious inference can still guide further investigation.
22. Scientific claims need the right expert source
A qualified scientist may have deep expertise in one field but not automatically in every subject. A published study may also be preliminary, limited or later challenged. Students should recognise the difference between a researcher explaining direct findings and a social-media personality extrapolating them to an unrelated policy.
Seek authoritative scientific organisations, current research summaries and the original study when the stakes warrant it. Check what participants or conditions were examined and whether conclusions were reported accurately. For personal health decisions, consult appropriate professionals rather than treating online material as individual medical advice.
23. Official policy claims require primary confirmation
Rules about school posting, subject levels, examinations or admission may be updated. Students and parents can encounter screenshots of authentic old documents with captions that imply they apply today. The reliable action is to check current MOE, SEAB or school communications for the relevant year and cohort.
Create a fictional case: a forwarded image claims a new Secondary 3 subject requirement. Ask the learner to search the school’s official handbook and the relevant authority rather than forwarding the image to other parents. Source verification can prevent unnecessary anxiety as well as misinformation.
24. A rumour about a classmate is not research material
Students may receive unverified stories about another young person’s behaviour, family or health. Attempting to investigate by sharing the allegation can increase harm even when the original claim is false. Information literacy includes knowing when not to circulate private material.
Encourage learners to avoid forwarding rumours, screenshots or personal images without consent. Where a genuine safety concern exists, use trusted school adults and appropriate reporting channels instead of public speculation. Respect for privacy is part of responsible evidence handling.
25. Disagreement does not prove one side is lying
People may disagree because they use different evidence, define terms differently or place weight on different values. Students should examine factual claims separately from judgments about what should be done. A debate can contain legitimate differences of preference alongside claims that are demonstrably incorrect.
Ask each speaker’s position to be restated fairly, then identify which parts can be tested against evidence. Do not assume that treating people respectfully requires treating every claim as equally well supported. The goal is fairness without false equivalence.
26. Uncertainty should be stated, not hidden
Some questions cannot be answered confidently from the available information. A student may feel that an assignment expects a definite statement, but guessing and disguising uncertainty undermines learning. Distinguishing confirmed facts, tentative inferences and unknown details is a sign of mature reasoning.
Use language proportionate to evidence: ‘The record confirms’, ‘The source suggests’, or ‘The available information does not establish’. Encourage students to specify what further evidence would improve confidence. An accurate uncertainty statement can be more useful than an impressive unsupported conclusion.
27. Citation is an evidence trail
A citation helps a reader inspect the source supporting a claim. It is not an ornament pasted at the end of a paragraph. Students should cite the actual document, data table or passage used, and distinguish direct quotation from paraphrase. A citation to an unrelated homepage does not verify a precise assertion.
Teach students to open each source before including it in their bibliography. Record title, author or organisation, date and relevant location where available. The goal is traceability: someone else should be able to check what the student used without guessing.
28. Quotation can be accurate but misleading
A brief quotation may reproduce the exact words while omitting a condition or qualification that changes meaning. The fact that quotation marks are correct does not guarantee the interpretation is fair. A selective sentence can invert the thrust of a longer statement.
Ask the learner to read the sentence before and after the quotation. Then restate the author’s argument in context. Good citation practice is both accurate transcription and faithful representation of the source’s meaning.
29. Search terms shape the answers you find
A student who searches ‘proof that this method is useless’ may receive different results from a student searching for the evidence on its effectiveness and limitations. Search terms encode assumptions and can narrow the field before evaluation begins.
Try writing one neutral search question and two competing hypotheses. Compare the sources each query finds. The learner should avoid choosing only results that confirm the preferred conclusion. Research quality begins with asking a question that permits a genuine answer.
30. Cross-checking does not require a hundred tabs
Endless browsing can create the appearance of diligence without improving the conclusion. A practical verification process selects a small number of sources with relevant expertise, independence and access to original evidence. The number of tabs is not a measure of epistemic quality.
Ask what one additional source would most efficiently resolve the main uncertainty. For a school deadline, the official school notice may be decisive. For a scientific claim, a study and an authoritative synthesis may be more useful than many derivative posts. Research should be targeted.
31. Media creation carries responsibilities too
Students are not only consumers; they may publish videos, slides, posters or class reports. They should check claims before posting, represent other people’s words fairly and identify image sources or AI assistance when required. Creative freedom and factual responsibility can coexist.
Before publishing a school project, review whether each important factual statement has support and whether any person appears without suitable permission. Explain why a dramatic but misleading visual could harm the audience’s understanding. Producing responsible media is a stronger test of literacy than merely spotting flaws in others’ work.
32. Correcting an error is a sign of integrity
Even careful learners can make mistakes. Responsible sharing includes acknowledging a wrong claim and correcting it where the original audience might have seen it. Quietly deleting a post may stop further spread but not repair an earlier misunderstanding among people who relied on it.
Teach a short correction: identify the original claim, state what was wrong, provide the verified information and explain where it came from. Avoid humiliating classmates who revise a view. An environment that rewards honest correction is more likely to produce accurate work.
33. Different G levels can use the same verification routine
At an introductory level, students may check the source, date and specific claim. A more demanding task might compare independent corroboration, sample limitations and alternative explanations. The cognitive principle remains similar, but the depth of analysis should fit the learner’s actual subject and development.
Do not describe a child as incapable of critical thinking because they study one subject at G1. Teach the next attainable step, provide clear examples and gradually increase independence. Equally, do not assume a strong G3 examination score makes a teenager immune to persuasive misinformation.
34. Application in History and Social Studies
Historical and civic questions ask students to interpret evidence from different perspectives. Provenance, purpose, context and corroboration affect what a source can establish. These disciplines are natural settings for media literacy, but their specific syllabus methods must still be taught rather than replaced by a generic checklist.
Ask whether a source is genuine, what it reveals about its creator, and how far it supports the given question. An opinion piece may be useful evidence of a viewpoint without proving every claimed fact. Precise interpretation matters more than automatically labelling a source trustworthy or biased.
35. Application in Science and Mathematics
Science requires evidence tied to models and methods, while Mathematics helps evaluate numerical claims. A viral chart that lacks axes or a headline that omits its denominator can be criticised using these subject skills. But mathematical correctness alone does not establish whether the sample represents the population.
Have the learner verify units, data source, collection method and inference. Then compare the claim with the underlying scientific question. The result is a cautious conclusion backed by actual evidence, not just a technically accurate computation.
36. A fictional school closure rumour
Fictional student Hana sees a forwarded message claiming that a school will be closed the following week. The screenshot has a familiar logo but no original link or date. Several classmates repeat it. Hana checks the school’s known communication channel and finds no such notice.
She refrains from sharing the claim and asks a responsible school adult for confirmation. The lesson is not that every rumour is false. It is that repetition, image design and emotional urgency do not replace a primary source when a decision about attendance is at stake.
37. A fictional manipulated comparison chart
Fictional student Amir sees a social post claiming that one learning method doubled success. The underlying chart compares a tiny group using one set of tasks with a much larger group using another. The comparison is not fair enough to support the dramatic headline.
Amir writes a more defensible summary describing the observed figures and their limitations. He asks what controlled comparison would be useful next. The act of revising a claim to fit the evidence is a practical demonstration of media and statistical literacy.
38. A fictional AI citation mistake
Fictional student Priya asks a chatbot for the date of a school policy change. It returns a precise date and a reference that looks official. When she opens the reference, it concerns an unrelated programme. She initially considers keeping the claim because the prose sounds confident.
Instead, Priya searches the relevant MOE announcement and corrects her notes. She retains the AI tool for brainstorming questions but no longer treats its reference list as verified automatically. The learning outcome is a better evidence workflow, not a ban on useful technology.
39. A six-week verification curriculum
Week 1 covers source, date and original context. Week 2 separates fact, inference and opinion. Week 3 practises independent corroboration. Week 4 examines charts, images and misleading presentation. Week 5 introduces AI-generated text, video and deepfake verification. Week 6 asks students to produce an accurate short report with sources and a correction protocol.
This is an illustrative learning plan, not a prescribed MOE sequence or SEC paper. It can be simplified or extended to the student’s actual level. The final test is a new claim: can the learner choose the right checking steps, explain the evidence and refrain from overclaiming?
40. The sixty-second decision to share
Before forwarding an alarming message, ask whether the source is identifiable, the claim is specific, the date is current and an independent primary source confirms it. Consider whether sharing might expose private information or cause unnecessary fear. The most responsible response may be to wait or verify.
Information literacy is not a demand for endless scepticism. It is a method for reaching proportionate confidence. Young people can learn to enjoy media, explore ideas and participate in civic life while refusing to confuse virality with truth.
Frequently asked questions
What does NLB’s S.U.R.E. stand for?
Source, Understand, Research and Evaluate, the National Library Board’s information-literacy framework.
Does a professional-looking site guarantee authenticity?
No. Check the domain, provenance and whether the claim exists on an independently reached official source.
Can a genuine photograph still mislead?
Yes. Cropping, timing and missing context can change the interpretation without digitally modifying the image.
Can a realistic video be a deepfake?
Yes. Important claims should be confirmed through trusted independent channels rather than appearance alone.
Are AI-generated citations always reliable?
No. Open each source and check that it exists and supports the specific claim.
Is every biased source useless?
No. A source may be useful for understanding someone’s viewpoint while remaining limited as evidence for another factual claim.
Does a high number of shares prove a claim?
No. Virality and repetition do not establish independent corroboration.
What is lateral reading?
Investigating a source using independent information outside the original page, rather than judging it solely by its design.
How should a student report uncertainty?
State what is confirmed, what is inferred and which information remains missing.
Does information literacy belong only to English or Humanities?
No. It also supports Science, Mathematics, digital tasks, project research and daily decisions.
What should students do with rumours about classmates?
Avoid forwarding or investigating publicly. Use appropriate trusted school contacts when a genuine safety concern exists.
What is the best verification habit?
Find the original claim, confirm its date and context, inspect independent relevant evidence and share only conclusions that evidence supports.
Official sources and eduKateSG pathways
The NLB S.U.R.E. school courseware explains Source, Understand, Research and Evaluate. MOE’s 2026 announcements describe AI literacy, information validation and deepfake awareness in Cyber Wellness. The Media Literacy Council provides national digital citizenship context. Follow the Full SBB master hub, digital learning and AI literacy guide, data literacy guide and academic reading guide for complementary study.
Closing: confidence should follow evidence
The strongest media-literate learner does not believe everything and does not dismiss everything. The learner can investigate the origin of a claim, read it carefully, distinguish observation from interpretation, seek credible corroboration and report an appropriately limited conclusion. Across G1, G2 and G3, those habits become a practical defence against error and a foundation for responsible participation in public life.