Information is a signal, difference or message that changes what a receiver can know, predict or do.
In one line: information works when something is encoded into a signal, travels through a channel, reaches a receiver, is interpreted in context, and reduces uncertainty enough to change understanding or action.
A page can contain thousands of words and still provide little useful information to a reader who cannot interpret the language. A single red warning light can carry enormous information to a pilot who knows what it means.
This tells us that information is not only about how much is transmitted. It is also about the relationship among signal, receiver, context and purpose.
What Is Information?
At a broad human level, information is something that changes uncertainty. A message tells us something we did not already know, confirms something uncertain, distinguishes among possibilities or directs attention toward what matters.
A useful information chain is:
Source → encoding → signal → channel → receiver → decoding → interpretation → update → action.
1. Information Begins With a Difference That Matters
If every possible state produced the same signal, the signal would tell us nothing. Information depends on distinguishable states.
A thermometer is informative because different temperatures produce different readings. An examination result is informative because different performances produce different evidence. A word is informative because it distinguishes one meaning or instruction from others.
The receiver needs some way to map the difference back to the world.
2. Encoding Converts Meaning Into a Transmittable Form
Thoughts do not travel directly between minds. They must be encoded into speech, writing, images, numbers, gestures, diagrams or another signal system.
Encoding always involves representation. A map encodes geography. A graph encodes numerical relationships. A grade compresses a complex performance into a small symbol. A headline compresses an event into a short phrase.
Compression makes transmission efficient, but it can also remove detail. The receiver should know what was lost.
3. Channels Carry Signals but Also Introduce Constraints
Information travels through channels: sound, paper, screens, radio, fibre, conversation, memory, databases and institutions.
Channels differ in speed, capacity, persistence and vulnerability to distortion. A face-to-face conversation carries tone and immediate feedback. A text message is persistent but may remove tone. A dashboard can compress thousands of records but may hide the individual cases beneath the aggregate.
The medium changes what is easy to transmit and what is easy to miss.
4. Noise Damages or Obscures the Signal
Noise is anything that interferes with the signal reaching or being interpreted by the receiver.
- literal background noise;
- poor handwriting or corrupted data;
- ambiguous wording;
- irrelevant detail;
- missing context;
- emotional overload;
- competing notifications;
- translation errors; or
- deliberate misinformation.
Reducing noise improves the probability that the received message resembles the intended one.
5. The Receiver Determines Whether the Signal Becomes Meaning
A signal does not interpret itself.
The receiver brings language, prior knowledge, expectations, goals and attention. The same sentence can be highly informative to an expert, confusing to a novice and misleading to someone who interprets a key term differently.
This is why communication must be designed for the receiver rather than only for the sender.
6. Context Changes Meaning
Many signals are under-specified without context. “It is cold” means something different in Singapore, a freezer and an Arctic expedition. “The score fell by ten points” means little without knowing the scale, baseline and conditions.
Context tells the receiver which interpretation is plausible and how much importance the signal deserves.
7. Relevance Determines Whether Information Is Useful
A person can be overwhelmed by accurate information that is irrelevant to the current job.
When a student is choosing what to revise tonight, the most useful information may be which topics repeatedly fail under retrieval — not the complete history of every mark earned since Primary school.
Information quality therefore depends partly on selection: which signal should reach this receiver for this decision now?
8. Verification Separates Information From Reliable Information
A message can be informative in the technical sense while still being false. A false rumour changes what someone believes even though it does not represent the world accurately.
Human information systems therefore need evidence checks: source tracing, corroboration, measurement quality, comparison with primary material and correction mechanisms.
UNESCO’s media and information literacy work emphasises the ability to access, analyse, evaluate, use and create information critically and responsibly — a vital capability in an environment of abundant sources, misinformation and AI-generated content.
9. Information Becomes Knowledge Only After Integration
Receiving a fact is not the same as understanding it. The new information has to connect with existing knowledge, survive checking and become organised enough to retrieve later.
Information is therefore an input to knowledge, not a synonym for knowledge.
This distinction becomes increasingly important in an age where access to information is cheap but judgement, integration and verification remain difficult.
10. Information Has Value When It Changes a Decision or Model
The value of information depends partly on what changes because it arrived.
A test result that confirms what everyone already knew may add little. One carefully chosen question that distinguishes between two competing diagnoses may add a great deal.
This is why the best next question is often the one that reduces uncertainty the most, not the one that collects the most data.
11. Information Systems Need Correction Paths
Errors are unavoidable in large information systems. The question is whether they can be detected and repaired.
Good systems preserve source, date, context and revision history. They distinguish fact from interpretation and make it possible to update a claim without destroying the entire knowledge structure.
The Whole Information Chain
Difference in the world → encode → transmit → survive noise → reach receiver → decode → interpret in context → verify → integrate with knowledge → update belief or action → preserve correction path.
A Useful Metaphor: Information Is a Package in Transit
The sender packages something for transport. The address determines the receiver. The channel carries it. Damage can occur in transit. The receiver opens it using expectations about what the symbols mean.
A package delivered perfectly to the wrong address has failed. So has a message transmitted flawlessly to a receiver who cannot decode it.
Successful information transfer is therefore end-to-end, not merely successful sending.
Information at Three Zoom Levels
Micro: the signal
What does this number, word, image or event distinguish?
Meso: the communication system
How do source, channel, context, receiver and feedback interact?
Macro: the information environment
Which institutions, platforms and norms determine what information is visible, trusted, amplified, corrected or forgotten?
How Information Fails
- Encoding failure: the signal does not represent the intended meaning clearly.
- Channel failure: information is lost, delayed or distorted in transit.
- Receiver mismatch: the message assumes knowledge or language the receiver does not possess.
- Context collapse: accurate fragments are separated from the conditions that give them meaning.
- Information overload: too many signals make the important distinction harder to detect.
- Misinformation: false content changes belief as though it represented the world.
- No correction path: outdated or false information continues circulating after better evidence exists.
How Information Is Repaired
Return to the source. Restore context. Reduce the message to the decision-relevant distinction. Check whether the receiver understands the code. Compare with independent evidence. Mark uncertainty. Preserve the date and provenance of the update.
Sometimes the repair is not more information. It is a better signal.
What Parents and Students Should Notice
- Who created this information?
- What was encoded, measured or omitted?
- What context is required to interpret it?
- Is it relevant to the question we are answering?
- Can the source be verified?
- Does it reduce uncertainty or merely increase volume?
- What should change if this information is true?
Information Quantity and Human Meaning Are Not the Same Thing
Information theory gives us a precise mathematical way to think about uncertainty reduction in signals. But a signal can carry a large amount of technical information while having almost no useful meaning for a particular human receiver.
A random-looking string may be highly unpredictable and therefore information-rich in a mathematical sense. Yet it may tell a parent nothing useful about a child’s learning. By contrast, one sentence such as “the learner understands the concept but repeatedly selects the wrong method under unfamiliar wording” may contain little raw data while changing the next educational action dramatically.
Signal information asks how uncertainty changes in the code. Human information asks what uncertainty changes for this receiver, about this world, for this job.
Entropy Describes Uncertainty Before the Signal Arrives
In information theory, entropy describes uncertainty across possible states before the next signal is known. A highly predictable signal carries less surprise; a less predictable signal can carry more information when it arrives.
This helps explain why the same observation can be informative in one context and almost useless in another. If a student normally scores between 70 and 75, another score of 73 changes little. If the same student suddenly produces 35 under identical conditions, the new result demands attention because it is surprising relative to the existing model.
But surprise is not automatically importance. A surprising signal can be noise, error or an irrelevant anomaly. The next question is whether the signal should change the model.
Compression Trades Detail for Efficiency
Human systems compress constantly. A grade compresses many answers into one symbol. A headline compresses an event. A dashboard compresses thousands of records. A diagnosis label compresses many observations into a category. Compression makes coordination possible because no receiver can process the full world at once.
The danger is forgetting that compression is a lossy representation unless every relevant detail can be reconstructed. A score can tell us that performance changed without telling us why. A national average can reveal a population pattern while hiding which groups moved in opposite directions.
Every summary should carry an implicit question: what disappeared when we compressed this?
Redundancy Can Look Inefficient and Still Make a System Safer
Perfect efficiency is fragile when channels are noisy. Communication systems often use redundancy so that missing or damaged parts can be detected or reconstructed. Humans do the same thing: we repeat key instructions, use both words and diagrams, confirm an appointment in writing after speaking, or ask a learner to explain an idea in another form.
Redundancy becomes wasteful when it merely repeats the same weak signal. It becomes valuable when independent or complementary signals reduce the chance of silent failure.
This creates a useful distinction:
Duplicate volume repeats information. Functional redundancy protects meaning against loss.
Error Detection and Error Correction Are Different Jobs
A system may be able to notice that a message is damaged without knowing the correct replacement. Checksums in computing, contradictory records in databases, or a teacher noticing that a student’s explanation conflicts with their answer can all detect a problem.
Correction requires additional information: a clean copy, independent source, repeated measurement, rule, model or return to the world. This is why trustworthy information systems preserve provenance and alternative routes rather than merely flagging “something looks wrong”.
Information Has a Provenance Chain
Useful information should be traceable. Where did this number, quote, claim or image come from? Who observed the original event? Which instrument or process produced the record? What transformations occurred before the information reached us?
A copied claim can travel through dozens of webpages and appear widely corroborated while still tracing back to one weak origin. A model output can look authoritative while being downstream of stale, incomplete or biased inputs.
Source → acquisition → transformation → storage → retrieval → presentation → receiver.
Every transformation is a possible distortion point. Provenance makes those points inspectable.
Selection Determines Which Reality Becomes Visible
No information system displays everything. Sensors sample. Editors select. Search engines rank. Dashboards choose indicators. Teachers choose which student response to inspect. Attention itself is a selection mechanism.
Selection is necessary, but it creates a hidden power: what is not selected can become practically invisible. A school dashboard focused only on examination results may miss attendance collapse, widening subgroup gaps or strong learning that a narrow assessment does not capture.
The right question is not “Can we avoid selection?” We cannot. It is: what selection rule is being used, whose purpose does it serve, and what important state could disappear because of it?
Information Bottlenecks Can Be More Important Than Information Volume
A system can possess excellent information that never reaches the person who can act on it. A teacher notices a pattern but the handoff fails. A frontline worker sees a safety issue but reporting is cumbersome. A student knows they are confused but has no safe channel to ask.
This creates an organisational information problem:
Signal available → signal detected → signal selected → signal transmitted → authorised receiver reached → receiver understands → receiver can act.
The loss can occur at any arrow. More data does not solve a blocked route.
Latency Changes the Value of Information
Some information remains useful for years. Other information decays rapidly. A historical source can become more valuable as archives improve; a traffic warning may be useless ten minutes late.
Information systems therefore need a time dimension: when was this observed, when was it transmitted, when did the receiver see it, and when does it expire?
In education, feedback after an examination can still support long-term learning, but it may be too late to change that examination performance. Timeliness is part of fitness for purpose.
Information Can Be Correct Yet Misleading Through Denominator, Scale or Framing
A true statistic can mislead when the denominator is hidden, the time period changes, absolute and relative risk are confused, or a graph truncates its scale. The content is not necessarily false; the representation changes what the receiver is likely to infer.
This is why information literacy includes representation literacy. Ask not only whether a number is accurate but how its presentation shapes interpretation.
Information Value Depends on the Decision It Can Change
Information has operational value when it changes what we should believe or do. A highly detailed report may have almost no decision value if every plausible result leads to the same action. One cheap discriminating observation can have enormous value if different outcomes imply different routes.
A practical value-of-information check asks:
- What uncertainty is blocking the next action?
- What observation could reduce it?
- Would a different result actually change the decision?
- How much does obtaining the information cost?
- What is the cost of delay?
AI Changes the Information Problem From Scarcity Toward Selection and Verification
Search engines, large language models and automated systems make it easier to produce, retrieve and transform information. That lowers one cost while increasing another: deciding which output deserves attention and trust.
An AI-generated answer is not automatically a source. It is a transformation whose inputs, retrieval path and reasoning may be partly hidden. For low-stakes exploration, that can still be useful. For consequential claims, the receiver should move outward toward primary evidence, authoritative records or independently verifiable receipts.
The human role therefore shifts from “find any information” toward frame the question → retrieve → compare → verify provenance → judge relevance → decide → observe consequences.
A High-Resolution Information Audit
- World state: What underlying difference are we trying to know?
- Acquisition: How was the signal detected or measured?
- Encoding: What representation was chosen?
- Compression: What detail disappeared?
- Channel: What can be lost, delayed or distorted in transit?
- Redundancy: What protects against silent loss?
- Provenance: Can the route back to the original source be reconstructed?
- Selection: Why was this signal surfaced while others were omitted?
- Receiver: What knowledge and context are required to decode it?
- Relevance: Which uncertainty does it actually reduce?
- Latency: Is the information still timely?
- Representation: Could framing, scale or denominator produce a misleading inference?
- Verification: What independent evidence checks the signal?
- Integration: Does the information connect to existing knowledge?
- Decision value: What action could reasonably change because this arrived?
- Correction path: How will later evidence update or replace it?
Evidence Boundary: Information Theory Does Not by Itself Define Meaning or Truth
Claude Shannon’s information theory gives a rigorous mathematical framework for communication under uncertainty, encoding, channel capacity and noise. It deliberately does not solve the full problem of human semantics or truth. UNESCO’s Media and Information Literacy work addresses the human side: access, evaluation, responsible use, misinformation and digital information environments.
The useful synthesis is therefore two-layered: information theory helps explain whether signals can survive transmission; evidence, language, knowledge and judgement help explain whether the resulting human interpretation deserves to guide belief or action.
Connect Information to the Wider eduKateSG Mechanism Estate
- How Language Works — how human communities encode meaning into shared symbolic systems.
- How Communication Works — how information becomes an end-to-end coordination problem between receivers.
- How Evidence Works — why a signal must be related to a claim before it can justify belief.
- How Knowledge Works — how verified information becomes integrated and retrievable.
- How Decision-Making Works — where information value becomes action value.
Causal Gateway Handoff
- How the World Works — place information inside the wider causal map.
- How Signal Systems Work — move upstream from meaning into detection, noise, encoding and receiver evidence.
- How Control Systems Work — follow information into state estimation, decision and correction.
- How Supply Chains Work — see information coordinate material, production, inventory and delivery across organisations.
Continue Through eduKateSG
Evidence and Further Reading
UNESCO’s Media and Information Literacy programme frames modern information competence around critically accessing, analysing, evaluating and using information, with particular attention to misinformation, digital technologies and AI. UNESCO also describes information literacy as the ability to seek, evaluate, use and create information effectively for personal, educational, occupational and social goals.
Frequently Asked Questions
Is information the same as data?
Not exactly. Data are recorded observations or symbols. They become informative when a receiver can interpret what differences they represent and use them to reduce uncertainty.
Is more information always better?
No. Irrelevant or badly organised information can increase cognitive load and obscure the signal that matters. Relevance and timing are part of information quality.
Can false information still influence people?
Yes. A false message can alter belief or behaviour. That is why verification and source evaluation are essential additions to mere transmission.
Final compression: Information works when a meaningful difference survives encoding, transmission, noise and interpretation well enough to reduce uncertainty — and when verification keeps the resulting update connected to the world.