Launching your learning experience…
Detailed notes on Automated and emerging technologies for Cambridge IGCSE Computer Science, covering key concepts, explanations, examples, and exam-focused revision points.
Computers performing tasks that normally need human intelligence — recognising images, understanding speech, making recommendations, diagnosing conditions. The components of an AI system, the structure of an EXPERT SYSTEM, what machine learning is and the trade-offs of trusting algorithms with consequential decisions.
Mapped to the Cambridge IGCSE 0478 syllabus (2026-2028).
Computer systems that simulate human-like reasoning — recognise, decide, learn, adapt.
Artificial Intelligence (AI) is the simulation of HUMAN-LIKE INTELLIGENCE by computer systems. The hallmarks:
Components of an AI system.
AI ≠ robot. AI is the SOFTWARE / decision-making side. Robots are physical machines. AI runs in your spam filter, Netflix recommendations, Google Maps, voice assistants — none of those need a body.
Cambridge tip. Mark scheme expects 'simulates human-like intelligence' AND at least two named components. Pure 'AI is intelligent computers' answers score poorly.
Captures human-expert knowledge in rules; uses it to give specialist advice.
An expert system is a kind of AI that captures the knowledge of HUMAN EXPERTS in a specific domain (medicine, law, engineering, geology) and uses it to give advice or diagnosis.
The four components.
1. Knowledge base. A large database of facts and rules elicited from human experts. For medical diagnosis: symptoms, conditions, test results, drug interactions, the relationships between them. Typically encoded as IF-THEN rules.
2. Inference engine. The reasoning component. Applies the rules in the knowledge base to the user's inputs to draw conclusions. Uses forward-chaining (start from facts, derive conclusions) or backward-chaining (start from goal, work back).
3. User interface. Lets the user enter information (symptoms, observations, history) and read the system's conclusions and recommendations. Usually question-and-answer style.
4. Explanation system. Tells the user WHY a particular conclusion was reached — which rules fired, which facts mattered. Crucial for medical, legal, financial uses where blind trust isn't acceptable.
Famous examples. MYCIN (medical diagnosis, 1970s). DENDRAL (chemistry). Modern expert systems power some clinical decision-support tools, tax-prep software and engineering diagnostics.
Cambridge tip. Mark schemes expect ALL FOUR components. The 'explanation system' is the one most commonly omitted — and the one Cambridge marks most heavily.
AI that learns patterns from data instead of being given rules.
Machine learning (ML) is a different approach to AI. Instead of being PROGRAMMED with explicit rules, an ML system LEARNS patterns from large amounts of training data.
The basic recipe.
Where ML wins. Tasks where the rules are too complex or vague for humans to write down — image recognition, language translation, voice transcription, recommending what to watch.
Where ML struggles.
Cambridge tip. Don't conflate AI and ML. AI is the umbrella; ML is one approach to AI; expert systems are another (rule-based) approach.
Powerful at scale — but raises bias, accountability and job concerns.
Advantages.
Disadvantages.
Cambridge tip. Top-band evaluation answers acknowledge BOTH the productivity gains AND the bias/jobs/accountability costs. One-sided answers (either techno-utopian or AI-as-existential-threat) lose marks.
Verbatim phrases and definitions Cambridge mark schemes credit.
AI appears on most Paper 1s, with growing weight in 2026-2028. Most-tested questions: define AI (4 marks), expert system components (8 marks), ML vs expert system (4-6 marks), advantages and disadvantages (6 marks). Examiner reports flag the 'omitted explanation system' error and conflation of AI with robotics.
Sources: Cambridge IGCSE Computer Science 0478 syllabus (2026-2028); 0478 Examiner Reports 2022-2024; 0478/12 Oct/Nov 2024 question paper and mark scheme. Last reviewed 2026-05-09.
Step-by-step solutions to past-paper-style questions on artificial intelligence, written exactly the way a tutor would explain them at the board.
Question
State two characteristics that are typical of an artificial intelligence (AI) system. (2 marks)
Step-by-step solution
Step 1
Recall the 0478 AI cycle: an AI system COLLECTS and STORES data, REASONS over that data, ACTS on a decision, then LEARNS and ADAPTS from new data/feedback. Any two of these characteristics earn the marks.
Step 2
Write two as short, distinct points so the examiner can tick each one — do not merge them into one sentence.
Answer
Examiner tip
Two clearly separated characteristics from the collect/reason/act/learn-adapt cycle. One mark each.
Question
A student says 'a calculator is an AI because it does maths faster than a human'. State whether the student is correct and give one reason. (2 marks)
Step-by-step solution
Step 1
Decide: the student is INCORRECT (1 mark). A calculator only follows fixed instructions — it never learns or adapts.
Step 2
Justify with the defining feature of AI (1 mark): AI must LEARN/ADAPT and make decisions from data; speed alone is not intelligence.
Answer
Incorrect. A calculator only follows fixed, pre-set rules and cannot learn, adapt or make decisions from data, so it is not AI — being fast at a fixed task is not the same as intelligence.
Examiner tip
1 mark for the correct judgement, 1 mark for a reason tied to learning/adapting/decision-making rather than to speed.
Question
Identify the four main components of an artificial intelligence (AI) system and give a brief role for each. (4 marks)
Step-by-step solution
Step 1
Knowledge base / data (1 mark). The store of facts and data the system reasons about (text, images, sensor readings).
Step 2
Rules (1 mark). The conditions/relationships the system uses to interpret the data (e.g. IF-THEN rules).
Step 3
Reasoning (1 mark). The method that applies the rules to the data to draw a conclusion / make a decision.
Step 4
Learning (1 mark). The method that lets the system adapt and improve using feedback or new data.
Answer
Knowledge base (facts/data), rules (how to interpret data), reasoning (draws conclusions), learning (adapts from new data). One mark per correctly named component with its role.
Examiner tip
Marks are for the component PLUS its role. A bare list of nouns with no role is capped — pair each name with what it does.
Question
Explain what is meant by machine learning and how it differs from a traditional rule-based program. (3 marks)
Step-by-step solution
Step 1
Define machine learning (1 mark). An AI approach where the computer LEARNS patterns from training DATA instead of being given every rule explicitly.
Step 2
Improvement point (1 mark). The model's accuracy IMPROVES as it is given more / better data — it adapts.
Step 3
Contrast (1 mark). A traditional program follows FIXED rules written by a programmer and does not change its behaviour unless the code is rewritten.
Answer
Machine learning lets a computer learn patterns from training data and improve with more data, rather than being explicitly programmed with every rule; a traditional rule-based program only follows fixed rules a programmer wrote and cannot improve on its own.
Examiner tip
Reward the 'learns from data' idea, the 'improves with more data' idea, and a clear contrast with fixed-rule programming.
Question
Describe the components of an expert system and explain how it reaches a conclusion when diagnosing a medical condition. (6 marks)
Step-by-step solution
Step 1
Knowledge base + rule base (2 marks). A store of FACTS and a set of IF-THEN RULES captured from human experts (doctors) — symptoms, conditions, test results and how they relate.
Step 2
User interface (1 mark). Lets the patient/clinician enter information (symptoms, history) and read the output, usually as a question-and-answer dialogue.
Step 3
Inference engine (1 mark). The reasoning component that APPLIES the rules in the rule base to the entered facts.
Step 4
Reaching a conclusion (2 marks). The inference engine matches the inputs against the rules, fires the rules whose conditions are met (chaining results together), and outputs the most likely diagnosis with a probability/recommendation — typically with an explanation of which rules were used.
Answer
Components: knowledge base + rule base (expert facts and IF-THEN rules), user interface (enter symptoms / read output), inference engine (the reasoner). To reach a conclusion the user enters symptoms via the interface; the inference engine applies the rules to those facts, fires the matching rules, and outputs the most probable diagnosis with a recommendation. Used as a decision-SUPPORT tool, not a replacement for a clinician.
Examiner tip
Mark scheme wants the named components AND a sequenced description of HOW a conclusion is reached (apply rules → fire matches → output diagnosis). Listing components without the reasoning process caps the mark.
Question
Explain the difference between narrow AI and general AI, and discuss two implications of organisations using AI to make decisions. (6 marks)
Step-by-step solution
Step 1
Narrow AI (1 mark). AI designed for ONE specific task (e.g. spam filtering, image recognition, recommendations). It cannot work outside that task — this is the only kind in real use today.
Step 2
General AI (1 mark). A hypothetical AI that could perform ANY intellectual task a human can, transferring knowledge across domains. It does not yet exist.
Step 3
Positive implication (2 marks). Faster, consistent decisions at huge scale — processes vast data 24/7, finds patterns humans miss (fraud detection, medical screening), reducing cost and human error.
Step 4
Negative implication (2 marks). Bias in training data can produce unfair decisions (hiring, lending); accountability is unclear when AI is wrong; and roles may be displaced. Decisions can be hard to explain ('black box').
Answer
Narrow AI = built for one task only (all current AI). General AI = hypothetical, human-level across any task, does not yet exist. Implications: (+) fast, consistent, large-scale data-driven decisions 24/7; (−) bias from training data → unfair outcomes, unclear accountability and reduced human oversight.
Examiner tip
1 mark each for clear narrow/general definitions, then a balanced 2+2 discussion of implications. A one-sided answer (only benefits or only risks) cannot reach full marks.
High-scoring sample answers for artificial intelligence on the Cambridge IGCSE 0478 paper, with examiner-style notes mapping each response to the mark scheme and assessment objectives.
State what is meant by the term artificial intelligence (AI). [1]
Model answer
Software / a computer system that simulates human intelligence — it can make decisions and learn/adapt from data rather than only following fixed instructions.
Why this scores
1 mark (AO1). Award for the simulation-of-human-intelligence idea OR the learn/adapt-and-decide idea — one clear correct point is enough.
State two characteristics of an AI system. [2]
Model answer
It collects, stores and reasons over data to make a decision / act on it.
It learns and adapts from new data or feedback, improving over time.
Why this scores
2 marks (AO1), 1 per characteristic. Any two of collect-and-store data, reason, act, learn/adapt. Two must be distinct points.
Identify three components of an expert system. [3]
Model answer
Knowledge base (and rule base) — the stored facts and IF-THEN rules.
Inference engine — applies the rules to reach a conclusion.
User interface — used to enter data and read the output. (Also accept: rule base as a separate component.)
Why this scores
3 marks (AO1), 1 per correctly named component. Accept knowledge base, rule base, inference engine, user interface — any three.
Explain one difference between machine learning and an expert system. [4]
Model answer
A machine-learning system LEARNS its own patterns from large amounts of training data (1) and IMPROVES automatically as more data is supplied (1). An expert system instead relies on FIXED rules and facts entered by human experts (1) and only changes when those rules are manually updated — it does not learn from new data on its own (1).
Why this scores
4 marks (AO1/AO2). Two linked points about machine learning (learns from data; improves automatically) and two about the expert system (human-written fixed rules; does not self-improve). Reward a clear contrast.
Describe how an expert system reaches a conclusion. [5]
Model answer
The user enters data (e.g. symptoms / answers) through the user interface.
The inference engine takes these inputs and searches the knowledge base / rule base.
It applies the IF-THEN rules to the facts, firing each rule whose conditions are met.
Results are chained together (the output of one rule feeds the next) until no more rules apply.
The system outputs the most likely conclusion/recommendation, often with a probability and an explanation of which rules were used.
Why this scores
5 marks (AO1/AO2). One mark each for: input via interface; inference engine queries knowledge/rule base; applies/fires matching rules; chains results; outputs conclusion (with probability/explanation). A sequenced process is required, not just a component list.
Explain the characteristics of artificial intelligence (AI). [6]
Model answer
Collects and stores data — an AI system gathers large amounts of input data (text, images, sensor readings) to work from.
Reasons over the data — it applies rules/patterns to interpret the data rather than just storing it.
Acts / makes decisions — it uses that reasoning to choose an output or action (e.g. a recommendation, a classification, a diagnosis).
Learns and adapts — it improves its decisions over time using feedback and new data, so its accuracy increases with experience.
Together these mean AI can handle tasks that normally need human intelligence, behaving differently from a fixed rule-based program because it changes its behaviour as it learns.
Why this scores
6 marks (AO1/AO2). Award for the four core characteristics — collect/store data, reason, act/decide, learn/adapt — each with a brief explanation, plus a linking point that distinguishes AI from a fixed-rule program. Mark up to 6 from the development given.
Definitions to memorise and the exact keywords mark schemes credit for artificial intelligence answers — sharpened from recent examiner reports for the 2026 0478 sitting.
The simulation of human-like intelligence by computers — including pattern recognition, decision-making, language understanding, learning and adaptation.
An AI system that uses a knowledge base of facts and rules from human experts to give diagnoses or recommendations in a specific domain.
The store of facts and rules an expert system uses. Built by capturing knowledge from human experts in the field.
The component of an expert system that applies the rules in the knowledge base to the user's inputs to draw conclusions.
The expert-system component that explains WHY a particular conclusion was reached (which rules fired). Essential for transparency.
An AI approach where the computer learns patterns from DATA rather than being given explicit rules. The model improves as more data is provided.
The data used to TEACH a machine-learning model. Quality and bias of training data directly affect the model's accuracy and fairness.
The traps other students keep falling into on artificial intelligence questions — taken from recent Cambridge IGCSE 0478 examiner reports and mark schemes — and how to avoid them.
Why it happens
Pop culture conflates the two.
How to avoid it
AI is the SOFTWARE / decision-making side. Robots are physical machines. AI can run with no robot involved (Netflix recommendations, spam filters, voice assistants on your phone).
0478 Examiner Reports 2022-2024
Why it happens
Less famous than knowledge base / inference engine.
How to avoid it
Mark scheme expects FOUR components. Memorise: knowledge base, inference engine, USER interface, EXPLANATION system.
Why it happens
Marketing hype.
How to avoid it
AI inherits BIAS from training data, can hallucinate plausible-sounding nonsense, and can't reason outside its training distribution. Mark scheme rewards balanced evaluation.
Why it happens
Both popular, both vague in the press.
How to avoid it
AI is the BIG umbrella. ML is ONE approach to AI (learning from data). Expert systems are another approach (rules). Don't equate them.
The things students keep getting wrong in this sub-topic, answered.