AI vs machine learning vs deep learning: how they fit together
Artificial intelligence is the broad field. Machine learning is an approach within AI that learns patterns from data. Deep learning is an approach within machine learning that uses multiple layers of learned representations. These are related ideas, rather than three competing products.
SchoolWhool's current learning path covers Python, Data Tools, Maths, Statistics and classical Machine Learning. Open lessons' parts and key terms are free to read.
The difference at a glance
One support inbox, three possible approaches
Suppose a small shop needs to route incoming messages to delivery, billing or returns. A hand-written rule can send a message containing an order-tracking phrase to delivery. That is a rule-based approach; whether such a simple automation deserves the AI label matters less than whether it solves the task.
A machine learning alternative uses historical messages with correct categories to fit a text classifier. A deep learning alternative learns a layered representation of those messages. Both require examples that reflect the messages the shop actually receives. Neither makes the categories themselves correct or useful.
Writing a reply is a different task from routing the message. A generative model could draft the reply after routing, but a plausible sentence about a refund must still be checked against the shop's policy and the customer's order.
Where computer vision, NLP and data science belong
Computer vision and natural language processing name problem areas: working with visual information and human language. Each can use rules, classical machine learning or deep learning. Image classification and text classification differ in their inputs, but both can be supervised learning problems.
Data science is a broader working practice around questions, data and evidence. An analysis of delivery delays may need careful grouping and a chart rather than a predictive model. Choosing a field name does not replace defining the question.
Choose an approach by testing the simplest useful solution
Write down the output
For the inbox, name the categories and decide what happens when a message fits two of them. Keep the labelling guide alongside the data.
Make a baseline
Try a clear rule or a simple classifier. Record the messages it gets wrong before adding a larger model.
Test the same cases
Compare approaches on held-out messages with the same labels and evaluation rules. Inspect rare categories and ambiguous wording separately.
Include the operating cost
Consider latency, data handling, maintenance and the consequences of a wrong answer. A small improvement on one metric may not justify a more complex system.
What should a beginner learn first?
If you cannot yet load a file and write a function, begin with Python. Next, use arrays and tables to inspect data, learn the statistics behind fair comparisons, and build a classical machine learning model with a held-out evaluation. That experience makes neural networks easier to reason about.
SchoolWhool's five published tracks cover that foundation. The course learning path combines them so you can study the maths and statistics alongside the modelling tasks that use them. It is a starting point for specialisation, not a claim that every AI field is already a SchoolWhool course.
Where this guide takes you
- Use the examples to understand the task and its evaluation. They are suggested learning exercises, not measured results or production-ready systems.
- The linked SchoolWhool courses teach the foundations. Specialist deep learning, computer vision, NLP, generative AI and reinforcement learning tracks are not published.
- Open lessons' parts and key terms are free to read. A free account opens quizzes and saved progress, plus practice and other material in the free sample modules.
Questions
Is all machine learning AI?
Machine learning is generally treated as a part of AI. The useful distinction is that ML fits behaviour from data, while AI also includes methods based on search, rules and explicit reasoning.
Is all AI deep learning?
No. Deep learning is one approach within machine learning, which is itself part of the wider AI field.
Is ChatGPT AI or machine learning?
Both descriptions apply at different levels. A language-model assistant is an AI application whose underlying model uses machine learning and deep neural networks.
Should I learn machine learning before deep learning?
Learn data preparation, supervised learning, loss functions and evaluation first. You can begin exploring small neural networks alongside that work without finishing every classical algorithm.
Does a data science project need AI?
No. A well-supported descriptive analysis can answer a question without training a model. Add prediction only when the question calls for it.
Sources
Technical references and further reading from the authors, educators and maintainers below. Read dates record when these references were checked; the examples and suggested exercises in this guide are SchoolWhool's own.
- Goodfellow, Bengio and Courville: Deep Learning, introductionRead 2026-10-14
- Google: What is machine learning?Read 2026-10-14
Related reading
What is AI?What is machine learning?What is deep learning?Machine learning vs deep learning: choosing a methodBuild the foundations for this subject
Follow SchoolWhool's learning path through Python, Data Tools, Maths, Statistics and Machine Learning. Start with the skills you need, read open lessons without an account, then sign in free for quizzes and saved progress. Free sample modules also open their practice and study material.
