Lesson 1 of 5 · 5 min

What AI Is (and What It Isn’t)

Learn what chat assistants and image generators really do, what they are good at, why they sometimes make things up, and whether you need to pay.

In this lesson

  • Tell the difference between a chat assistant and an image generator
  • Know the tasks AI handles well and where it struggles
  • Understand why AI sometimes gives confident wrong answers

What we mean by “AI”

When people talk about AI today, they usually mean a computer program that understands everyday language and creates something new: a reply, a summary, a picture, even a short song. You type a request in normal words, and the program answers in seconds.

These tools learned by studying a huge amount of text and images. They did not learn facts the way a student does. Instead, they learned patterns: which words usually follow other words, and which shapes and colours usually appear together. When you ask a question, the AI builds an answer piece by piece, choosing what is most likely to fit.

As a beginner, you will mostly meet two kinds of tool:

  • Chat assistants, such as ChatGPT, Gemini, Copilot and Claude. You talk to them in a chat window. They write, explain, summarise, translate, plan and answer questions.
  • Image generators, such as Midjourney, Ideogram or the image features inside some chat assistants. You describe a picture in words and they create it.

Some tools now do both, and many can also read a photo or a document you upload. The basic idea stays the same: you describe what you want, and the AI produces a first version.

What AI is good at, and where it struggles

AI assistants are strongest at tasks with words and ideas. Think of a fast, helpful assistant who has read a lot of books but has never left the office. They are very good at:

  • Drafting emails, messages, short texts and outlines.
  • Rewriting text so it is shorter, simpler, friendlier or more formal.
  • Explaining a difficult topic in plain language, with examples.
  • Brainstorming ten fresh ideas when you feel stuck.

They struggle with other things. They may not know very recent news unless they can search the web. They can make mistakes with maths and counting. They know nothing about your personal situation unless you tell them. And sometimes they simply make things up.

Why AI makes things up

Because an AI predicts likely words, it does not always check whether those words are true. If it does not know the answer, it may still produce one that sounds correct. People call this a “hallucination”. It can be a fake statistic, a book that does not exist, or a quote that nobody ever said.

Here is a simple before and after.

Before: “Who won the town chess championship in my city in 2019?” The AI gives a name, a score and a nice story. It all sounds real, but you have no idea where it came from.

After: “Who won the town chess championship in my city in 2019? If you are not sure, say so clearly and tell me where I could check.” Now the AI is much more likely to admit it does not know and point you to an official source.

The lesson: treat AI answers as a helpful first draft, not as the final truth. The more important the answer, the more carefully you should check it.

Free or paid?

Most popular assistants have a free version that is more than enough for learning and everyday use. Paid plans usually give you more messages per day, faster replies, newer features and better handling of long documents or images.

A sensible approach is to start with a free version, use it regularly for a few weeks, and only think about paying if you keep reaching the limits. You do not need several subscriptions to get real value.

Remember that “free” does not mean “private”. Look at the privacy settings of any tool you use, and never type anything you would not want a stranger to read. We cover this in detail in the safety lesson.

Practice

Open any AI assistant you already use, such as ChatGPT, Gemini, Copilot or Claude. Paste the prompt below and replace the part in [BRACKETS] with your own idea.

Prompt
I am completely new to AI. Explain in simple words what an AI chat assistant is, using one example from [EVERYDAY_SITUATION]. Then list 3 things you are good at and 3 things you often get wrong. Keep the whole answer under 200 words.
[EVERYDAY_SITUATION]
e.g. cooking dinner, a school day, running a small shop
See an example result

Think of me as a very well-read helper who answers in text. I learned from lots of writing, so I can guess what a useful reply looks like.

Example from cooking dinner: you tell me you have rice, eggs, onions and spinach. I can suggest three simple recipes, write the steps, and adjust them if you say "no spicy food".

What I am good at:
1. Writing and rewriting texts, like messages or shopping lists.
2. Explaining things in simple words.
3. Giving ideas when you feel stuck.

What I often get wrong:
1. Exact facts, dates and numbers. I can sound sure and still be wrong.
2. Recent news, unless I can search the web.
3. Your personal details. I only know what you tell me.

Tip: always check important information with a trusted source.

What to notice: A good answer admits its own limits clearly, not only its strengths. If the reply is too long or too technical, ask a follow-up such as "Say it again in 80 words for a 12-year-old".

Key takeaways

  • AI predicts likely words and images; it does not truly know facts
  • It is great for drafting, rewriting, explaining and brainstorming
  • Free versions are enough to start; check anything important