Part 2 · How AI Actually Works
What AI Actually Is (Not Magic, Not Movies)
It Felt Like Mind-Reading
The last time your phone finished your sentence for you, it felt like mind-reading. You typed "running", and it offered "late". For a second it seemed the phone had read your mind. It had not. It had seen the word "running" followed by "late" a few million times, and it played the odds.
Why This Matters
Most people meet AI as a headline, either a miracle or a menace. Neither view helps you decide anything on a Monday morning. Before you can use AI well, you have to see it plainly, without the movie soundtrack. By the end of this lesson you will be able to:
- Explain what AI is, in plain language a colleague would follow
- Recognise three everyday tools that are already AI
- Describe why AI is just maths, and why that changes how you use it
This is Understanding AI, the first dimension of your personal AI readiness.
What AI Actually Is
For decades, software worked one way: a person wrote the rules. If temperature reads above thirty-eight, flag a fever; every step spelled out by a human in advance. AI flips this. Instead of being handed the rules, the machine is shown thousands of examples and works out the pattern itself. Show it enough spam, and it learns what spam looks like, without anyone writing a single rule.
| Traditional Software | Artificial Intelligence |
|---|---|
| Humans write the rules Every step is spelled out in advance. | The machine finds the rules Show it thousands of examples; it learns the pattern. |
| Predictable, but rigid It only does what it was told. | Flexible, but only as good as its data It knows nothing it has not seen. |
So where is the intelligence? A trained model is just a very large set of numbers. Training nudges those numbers, over and over, until the model's guesses match the known answers. When you use it, the model runs arithmetic on your input and returns the most likely output: no understanding, no intent. That is powerful, and it is also why AI can be confidently wrong the moment your question falls outside what it has seen.
- Your inbox separates spam from real mail: learned from billions of flagged messages
- Your streaming app queues the next show: learned from what similar viewers watched
- Your photos app groups every shot of the same face: learned from images
Key Takeaway
To use AI well you need one habit: when software impresses you, ask what it learned from.
Check your understanding
Self-check only. Nothing is scored and nothing is recorded; open an answer to compare it with your own.
Your team lead says the company's customer support chatbot 'uses AI to understand questions and write replies.' A colleague argues it is just automation because the answers feel scripted. Based on what you learned, which response best applies?
- It is automation, all chatbot responses are pre-written by humans.
- If the chatbot generates responses by matching language patterns from training data, it learned its behaviour from examples, even if the output sometimes feels formulaic.
- It is AI only if it gets every answer right.
- It is impossible to tell without seeing the source code.
Show the answer
If the chatbot generates responses by matching language patterns from training data, it learned its behaviour from examples, even if the output sometimes feels formulaic.
A finance colleague shares a news article claiming 'AI will replace forty per cent of jobs within five years.' She asks if she should be worried. Based on what you learned, which response best applies?
- Yes. AI is advancing so fast that most finance roles will be automated soon.
- No. AI is just maths, so it cannot do anything a spreadsheet formula cannot do.
- The useful response is to ask what the claim rests on, then sort her own tasks: which are pattern work a machine learned from examples, and which need her judgement.
- She should ignore it. AI predictions in the media are never accurate.
Show the answer
The useful response is to ask what the claim rests on, then sort her own tasks: which are pattern work a machine learned from examples, and which need her judgement.
Saved in this browser only. Carry progress between devices with the resume link on the course page.