AI Basics
What Is Artificial Intelligence?
A clear definition of AI: narrow vs general, rule-based vs learned systems, everyday examples, and common myths.
Tutorial overview
What you will learn
- Define artificial intelligence in one accurate sentence
- Distinguish narrow AI from general AI and rules from learning
- Debunk the most common AI misconceptions with examples
By the end, you will have
- A precise beginner definition of AI you can explain to someone else
Introduction
Artificial intelligence is the field of building computer systems that perform tasks which normally require human intelligence, such as understanding language, recognizing objects, making decisions, and learning from experience. That is the whole definition. AI is not a single machine, not a conscious mind, and not necessarily a chatbot. It is a broad family of techniques for making software behave capably in situations that are too messy for simple fixed instructions.
This lesson unpacks that definition so you can use the term precisely and spot the myths.
What you will build or practice
You will write your own one-sentence definition of AI, sort a list of everyday technologies into AI and not-AI, and correct three common misconceptions in your own words. These are the fastest checks that you truly understand the concept.
Before you begin
You need a web browser and a notes app. No math or programming is required. If terms like model or training feel unfamiliar, keep the AI vocabulary cheat sheet open in another tab.
Key concept
Two distinctions do most of the work in understanding AI.
Narrow vs general. Every AI system deployed today is narrow: it does one kind of task, such as translating text, detecting spam, or generating images. General AI, a system with flexible human-level ability across most tasks, remains a research goal, not a product. When headlines blur this line, confusion follows.
Rule-based vs learned. Early AI was rule-based: programmers wrote explicit instructions, like a tax program's decision tree. Modern AI is mostly learned: the system is shown many examples and adjusts itself to find patterns, which is the machine learning approach. Learned systems handle messy inputs like photos and speech far better, but their mistakes are harder to predict.
Step 1: Write the definition in your own words
In your notes, complete this sentence: AI is software that... Aim for something like software that handles tasks needing judgment by learning patterns from data. If your sentence includes robots, consciousness, or thinking, revise it. Capability, not consciousness, is the test.
Step 2: Sort everyday examples
Classify each of these: a calculator, a spam filter, a spreadsheet formula, face unlock on a phone, a map app choosing a route, a chat assistant. The calculator and spreadsheet follow fixed instructions, so they are not AI. The spam filter, face unlock, route planner, and chat assistant all learned patterns from data, so they are. The dividing line is learned behavior on messy input.
Step 3: Identify the type
For each AI example from Step 2, note whether it mainly recognizes (face unlock, spam filter), decides (route planning), or generates (chat assistant writing text). Generation is the newest branch, covered in Generative AI Explained for Beginners.
Step 4: Correct three misconceptions
Write a one-line correction for each myth:
- Myth: AI understands what it says. Reality: it predicts plausible patterns without human-style comprehension.
- Myth: AI is objective because it is math. Reality: it inherits biases and gaps from its training data.
- Myth: AI is either useless hype or about to replace everyone. Reality: it is a capable, fallible tool whose value depends on how you use and verify it.
Practice exercise
Explain AI to a friend or family member who has never studied it. Use your definition, one everyday example, and one myth correction. If they can repeat the idea back accurately, you have passed the real test of understanding.
Common mistakes
- Using AI, machine learning, and chatbot as interchangeable words.
- Assuming anything impressive a computer does is AI.
- Judging all AI by one bad chatbot answer, or trusting all AI because of one good answer.
Check your understanding
- What makes a system AI rather than ordinary software?
- Why does the narrow vs general distinction matter when reading news?
- Give one example each of rule-based and learned software.
Key takeaways
- AI means software performing tasks that normally need human judgment.
- All current AI is narrow; general AI does not exist yet.
- Modern AI learns patterns from data, which explains both its power and its unpredictable errors.
Next steps
The learning part of the definition is machine learning, explained next in What Is Machine Learning? For a hands-on start, revisit AI Basics.
Related resources
Newsletter or next lesson
Join the newsletter for clear weekly lessons, or continue to What Is Machine Learning?