Start with the big idea
Think of AI as a field of problem-solving methods. Some systems follow rules written by people; others learn statistical patterns from data. Deep learning is one family of machine learning approaches. Generative AI describes a goal, creating outputs, rather than a guarantee about a single model architecture.
Understand it in more detail
Artificial intelligence is a broad area concerned with systems that perform tasks associated with perception, decision-making, prediction, language and reasoning. A fixed-rule chess program may be AI without learning its rules from examples.
Machine learning fits statistical models to data. A spam classifier may learn weights from labelled messages, while a traditional rules-based filter can operate without training. Deep learning uses multilayer neural networks to learn representations; it is a subset of machine learning.
Generative AI produces new samples, such as text or images. Many modern generative models use deep learning, but the categories are not a perfect set of nested circles: generation also has older statistical and rule-based approaches.
Worked example
A school timetable system that simply checks fixed conditions is automation. A trained model predicting how crowded the library will be is machine learning. A language model drafting an announcement is generative AI. Classify each by how it works, not how impressive it sounds.
A misconception to avoid
Not every AI system learns from data; not every ML system is a neural network; and not every output from a generative model is reliable or original in a copyright sense.
Check your understanding
Think first, then open each answer to check your explanation. These self-checks do not record a score.
1. A system applies 20 manually written rules. Must it be ML?
No. Fixed rules can automate decisions without a learned statistical model.
2. Does 'generative' mean 'always true'?
No. Generation is the ability to produce samples; truth requires verification.
Try it yourself
Draw a three-column comparison of rules, learned predictions and generated outputs. Choose an everyday technology for each and explain what evidence you would need to confirm your classification.
Explain it back: Summarize the main idea in your own words and name one limitation or risk.
Continue learning
Read an independent reference: https://www.ibm.com/think/topics/artificial-intelligence. This link is for reference, not a requirement to create an account. External websites may have different privacy and access policies.
Educational pilot. Terminology, age suitability and activities are subject to expert review. These pages do not send your answers to a server.