In simple terms
A friendly intro before the formal notes — no formulas yet.
AI: The Digital Recipe Book
Algorithms are like recipes that computers follow to solve problems or make decisions. Artificial intelligence uses complex algorithms, often learning from data, to mimic human-like intelligence.
Think of an algorithm as a recipe for baking a cake. The recipe lists specific ingredients (data) and a step-by-step method (the algorithm) to get a predictable result (the cake). AI is like a master chef who can invent new recipes (algorithms) by tasting thousands of cakes (learning from data) to create the perfect one for any occasion, sometimes even creating things no human chef has thought of.
- 1
Identify the Algorithm: First, pinpoint the specific set of rules or instructions being used to process information and make a decision. What is the 'recipe'?
- 2
Analyse the Data: Examine the data the algorithm is trained on or uses. Is it representative? Is it biased? What are the 'ingredients'?
- 3
Evaluate the Outcome: Assess the decision or output produced by the algorithm. Is it fair, accurate, and transparent? What is the 'cake' like?
- 4
Consider the Impact: Critically evaluate the broader societal, ethical, and economic consequences of using this algorithm in a real-world context. Who benefits and who is harmed?
Explore the concept
Use the live diagram and synced steps — play it or tap a step card to walk through.
Full topic notes
Formal explanation with the rigour you need for the exam.
Deconstructing the Digital Brain: Algorithms, AI, and Machine Learning
It's easy to use these terms interchangeably, but they have distinct meanings. An algorithm is the most fundamental concept; it's simply a set of instructions for completing a task. Artificial Intelligence (AI) is a much broader concept, representing the goal of creating machines that can simulate human intelligence. Machine Learning (ML) is a popular and powerful method for achieving AI. Instead of programming explicit rules, ML involves providing a system with vast amounts of data and allowing it to learn the rules for itself.
All Machine Learning is AI, but not all AI involves Machine Learning (some AI uses hand-coded rules, known as symbolic AI or 'Good Old-Fashioned AI').
All AI systems use algorithms, but not all algorithms are part of an AI system (e.g., a simple sorting algorithm is not AI).
Deep Learning is a specialised subset of Machine Learning that uses complex neural networks, driving recent breakthroughs in AI.
The Fuel for Intelligence: Data and Bias
A machine learning model is only as good as the data it is trained on. This principle is often summarised as 'Garbage In, Garbage Out' (GIGO). If the data used to train an AI reflects existing societal biases, the AI will not only learn these biases but can also amplify them at scale. For example, if an AI is trained on historical loan data where a certain demographic was unfairly denied loans, the AI will learn to associate that demographic with high risk, perpetuating the injustice.
The 'Black Box' Problem and Explainable AI (XAI)
Many of the most powerful AI models, particularly in deep learning, are 'black boxes'. We can see the input (e.g., a medical scan) and the output (e.g., 'cancer detected'), but the internal decision-making process is so complex that it's opaque even to its creators. This poses a significant problem for trust and accountability. How can we trust a medical diagnosis if the doctor can't explain why the AI made it? Who is responsible if a self-driving car makes a fatal error? Explainable AI (XAI) is an emerging field of research and practice that aims to develop techniques to make AI decisions understandable to humans, turning the black box into a more transparent 'glass box'.
Worked examples
See the formulas applied — reveal one step at a time, like the exam.
An AI-powered recruitment tool is trained on a company's hiring data from the last 20 years. The data shows that 80% of successful hires in technical roles were male. When deployed, the tool consistently ranks male candidates higher than female candidates with similar qualifications. Using the concept of algorithmic bias, explain this outcome and suggest two distinct mitigation strategies. [6 marks]
- 1
• Explanation of Bias (2 marks): The outcome is a clear example of algorithmic bias. The AI has learned a spurious correlation from the historical training data: that being male is a strong predictor of being a successful hire. It has codified the company's past, potentially biased, hiring practices, rather than identifying the best candidates based on skill alone. The algorithm is not 'sexist'; it is simply reflecting and amplifying the bias present in its training data.
A hospital uses a 'black box' deep learning model to predict the likelihood of a patient developing a specific type of cancer from medical scans. The model predicts a high risk for Patient A, but the doctors are unsure why. Evaluate the ethical implications of using this technology for patient diagnosis, referencing the concepts of transparency and accountability. [8 marks]
- 1
• Identification of Ethical Issues (2 marks): The primary ethical issues are the lack of transparency and accountability. Using a 'black box' model in a critical healthcare context raises concerns about patient autonomy, informed consent, and the potential for harm if the model is wrong and its reasoning cannot be scrutinised.
How it all connects
The big idea sits in the middle — tap a linked idea to explore the link.
Tap a linked idea to see how it connects back to the main topic — that connection is what examiners reward.
Glossary
Try to recall each definition before you reveal it.
Quick check
Answer in your head first — then tap to check. No pressure.
Revision flashcards
Flip the card. Test yourself before the exam.
Algorithm
A finite sequence of well-defined, computer-implementable instructions designed to solve a class of problems or to perform a computation. Think of it as a step-by-step recipe.
Key takeaways
Review these before you close the topic — retrieval beats re-reading.
- ✓
All Machine Learning is AI, but not all AI involves Machine Learning (some AI uses hand-coded rules, known as symbolic AI or 'Good Old-Fashioned AI').
- ✓
All AI systems use algorithms, but not all algorithms are part of an AI system (e.g., a simple sorting algorithm is not AI).
- ✓
Deep Learning is a specialised subset of Machine Learning that uses complex neural networks, driving recent breakthroughs in AI.
Practice — then mark it
The whole point: a real Cambridge question, marked mark-by-mark.
Practice Questions: Algorithms and AI
Practice Questions: Algorithms and AI
Extra simulations & links
PhET, GeoGebra and other curated tools — open in a new tab.
Frequently asked
Checkpoint
One marked question is worth ten re-reads — close the loop before you move on.
Reading it isn’t knowing it — prove it.
Before you move on: do Practice Questions: Algorithms and AI on paper, snap a photo, and get examiner-style feedback on exactly where you win and lose marks.