What Is AI, Really?
Understand what AI can do, how generative AI differs from other software, and why its answers can sound more certain than they are.
25 minutes Course overview1. AI is less mysterious than it sounds
Artificial intelligence (AI) is a broad name for computer systems designed to perform tasks that usually involve human judgement, such as recognising patterns, interpreting language or making predictions. Your spam filter, route-planning software and a chatbot are all very different uses of AI.
Generative AI is the family we’re exploring here. Rather than only sorting or identifying information, it can produce new material: words, pictures, audio, code, summaries and ideas. You describe what you want; the system generates a response.
Think of it as a powerful assistant for working with information, not an all-knowing authority. It can be helpful without being correct, and fluent without understanding your circumstances as another person would.
2. What happens when you type a message?
Your message is called a prompt. A language model processes the words and their relationships, using patterns learned during training and information supplied in the conversation to generate a response. You don’t need the mathematics to use it any more than you need to build a car engine to drive.
Most chat tools also have other capabilities or connect to tools, such as searching websites, reading documents or working with images. Those aren’t the same thing as the language model’s built-in knowledge. If a chat doesn’t actually have the relevant tool or permission, it cannot genuinely access your files, diary or the latest news just because you ask it to.
3. The three things to remember
- AI is good at first drafts and patterns. Ask it to simplify a difficult explanation, compare ideas, draft a polite message or brainstorm approaches.
- AI can invent details. It may state incorrect facts, quotations, laws, sources or calculations with complete confidence. We call this a hallucination.
- You’re the decision-maker. Use AI to think better and work faster. Keep responsibility for checking important information and choosing what to do.
4. Everyday examples
At home: turn “I have pasta, tomatoes and peppers” into three dinner ideas. At work: ask for a clearer agenda using fictional project details. When learning: request an explanation of interest rates as if you’re completely new to them.
Try it yourself
Open an AI assistant and ask the same question two ways. Try a topic you know a little about:
Compare the two responses. Which was easier to follow? Why? Write one sentence describing what AI can do and one sentence describing what it cannot guarantee.
Can you explain it back?
Answer each question before opening the explanation. You aren’t being graded — this is how you make the knowledge stick.
1. Does a confident AI answer mean it is correct?
No. Confident wording is not proof. Important facts still need checking.
2. What’s the difference between traditional and generative AI?
Traditional AI often classifies or predicts; generative AI creates new content such as text or images.
3. Can a chatbot open your files simply because you tell it to?
Not necessarily. It needs the relevant access, capability and permission.