Module 0 · Foundation · Free — start here

AI Fundamentals:
What Is It, Really?

Before we get into AI judgment, we need to build a shared foundation. No assumptions. No jargon. This module applies to every learner, every sector — it's where we all start.

By the end of this module, you will be able to:
  1. Explain the difference between AI and non-AI software using examples from your own work
  2. Name and describe the four types of AI you are most likely to encounter at work in South Africa
  3. Explain why AI outputs can sound authoritative and still be wrong
  4. Identify at least one AI system you already encounter at work and classify it by type
55 minutes · 4 phases All sectors · No technical background required Phases 1 & 2 free — register to unlock all phases
1 · Concept
2 · Encounter
3 · Reflect
4 · Apply
1
Concept — What AI actually is
The real explanation, without the hype · ~25 min
Lesson 1 of 7 — The Librarian Analogy

Before we begin

Every generation has had its version of this moment. ATMs were supposed to eliminate bank tellers. Computers were going to make office workers obsolete. The internet was certain to destroy retail and journalism. Smartphones were going to ruin attention spans. In every case, the anxiety was real — and not entirely wrong. Things changed. Jobs transformed. New skills became essential.

In each shift, the people who stayed in control were those who understood the technology — not the ones who feared it most, and not the ones who trusted it blindly. The ability to ask "what does this tool actually do, and when should I not rely on it?" made the difference.

AI is this generation's version of that moment. The tools are new. The pattern is familiar. This module gives you what you need to understand the tool — so you can shape how it changes your work, rather than having it happen to you.

Lesson 1

The Librarian Analogy

Imagine a librarian who has read every book in the library.

You ask: "How do I start a business?"

The librarian doesn't think. The librarian remembers patterns from all the books — what business books usually say when someone asks that question.

AI is similar

Except it has read far more than any library — and responds much faster. But like the librarian, it is recalling patterns, not reasoning from scratch.

Lesson 2

The World's Fastest Guesser

Imagine I show you this sentence:

The sun rises in the _____.

You immediately know the answer. Why? Because you've seen thousands of examples before. Now try this:

Happy birthday to _____.

You can probably guess: you, her, him. Again — you're predicting.

Now imagine you had read every book, every website, millions of conversations, billions of sentences. You'd become extraordinarily good at guessing what word comes next.

That is essentially what ChatGPT does

Not thinking. Not understanding. Predicting.

Try it yourself

Finish these sentences — write the first word that comes to mind:

Lesson 3

Why AI Feels Intelligent

Why does a parrot sound like it understands English?

Because it can repeat words.

Why does AI sound intelligent? Because it has seen enormous amounts of language and learned patterns.

Example

If I ask AI to write a resignation letter — it has seen millions of resignation letters. It predicts what one usually looks like.

But it doesn't know your boss, your feelings, or your company. It only knows patterns.

Lesson 4

Why AI Makes Stuff Up

Imagine I ask you: "Tell me what happened at my wedding."

You can't know. But some people — trying to be helpful — will guess anyway.

AI does the same thing. When it doesn't know, it predicts. And prediction sometimes creates fiction.

This is called Hallucination

Not because AI is broken or lying — because it is guessing. It completes the pattern even when it has no real information to draw on.

Try this yourself

Open ChatGPT, Copilot, or any AI and ask:

What were the three main findings of the 2023 Human Sciences Research Council report on AI use in South African workplaces?

It will likely produce three detailed, specific, confident findings — complete with percentages and recommendations. That report may not exist at all. AI is completing the pattern of what a research report usually sounds like, not retrieving a real document.

Lesson 5

AI Is Not One Thing

People think of AI as a robot, or a chatbot. But you're already using AI every day — you just didn't know it was AI.

🎵
Spotify
Predicts music you'll like
📺
Netflix
Predicts shows you'll watch
📧
Gmail
Predicts what's spam
📱
Face ID
Predicts whether it's your face
💳
Credit Score
Predicts whether you'll repay
🤖
ChatGPT
Predicts the next word

The key insight

AI is not a robot. AI is prediction applied to different problems. The tool changes. The core mechanism — learn patterns, make predictions — stays the same.

Lesson 6

Smart at Some Things, Helpless at Others

A dog can recognise faces, learn commands, and solve problems. But a dog doesn't understand mathematics.

AI is similar. It can be incredibly capable in some areas — and completely unreliable in others.

Where AI is powerful
  • Processing huge amounts of text quickly
  • Finding patterns in data humans would miss
  • Drafting, summarising, translating
  • Recognising images and faces at scale
Where AI fails
  • Knowing when it's wrong
  • Checking facts or verifying sources
  • Understanding your specific context
  • Taking moral responsibility for decisions
The mistake most people make: "If AI is smart at one thing, it must be smart at everything." That assumption is where most AI mistakes in the workplace happen.
Lesson 7 — The Most Important Rule

AI is a prediction engine,
not a truth engine.

Once you have this, everything else clicks into place.

Bias?
Prediction from historical patterns that contain discrimination
Hallucinations?
Prediction without enough real information to draw on
Recommendations?
Prediction of what you'll like, based on what others did
Credit scoring?
Prediction of repayment, based on patterns in past data

If you finish this section understanding only one thing — this is the one. Everything else in this course builds on it.

Knowledge Check 3 questions · answer all before moving on

1. What best describes how AI works?

2. A generative AI tool gives you a confident, detailed response. What should you do?

3. Spotify, Netflix, Gmail and Face ID all use AI. What do they have in common?

2
Encounter — Spot the AI
Now that you know what AI is — spot it in the wild · ~10 min

Below are 10 everyday tools and situations. Pick your gut answer: does this use AI, or not?

Click once to mark AI (green) · click again to mark Not AI (amber) · click a third time to clear.

📧
Gmail spam filter
Automatically puts junk email in your spam folder
AI ?
No human wrote a list of rules for every type of spam. The filter learned patterns from millions of flagged emails. Each time you correct it, it updates. That learning from examples — not rules — is the hallmark of AI.
🎵
Spotify Discover Weekly
New music playlist generated for you every Monday
AI ?
A recommendation model analysed your listening history and compared it to millions of users with similar taste to predict what you'd enjoy next. No human curated that playlist. Prediction from patterns across massive data is AI.
📊
Excel =SUM formula
Adds up the numbers in a column
Not AI ?
It executes an exact rule you gave it: add these numbers. It will never get "better" at summing no matter how many times you use it. No learning. No prediction. Software that follows fixed instructions without adapting is automation — not AI.
📱
Face ID / facial recognition
Your phone unlocking when it sees your face
AI ?
Trained on thousands of facial images to recognise structural patterns unique to your face. There is no rule like "if nose = X, unlock" — the system learned what your face looks like from examples. That's computer vision AI.
🔍
Ctrl+F search in a document
Find a specific word in a file
Not AI ?
Scans text character by character for an exact match. No interpretation, no understanding of meaning, no learning. Search "cat" and it won't find "kitten." The same input always gives the same output. Fully deterministic — the opposite of AI.
🛒
"You may also like" product suggestions
Shopping site recommendations based on browsing
AI ?
A recommendation algorithm detected patterns in your browsing, your purchases, and the behaviour of shoppers who bought similar things — then predicted what you'd likely buy next. The output changes as you change. That's AI adapting to you.
📄
A PDF form you fill in manually
Digital form with fixed fields — no automation
Not AI ?
A static digital document. No processing happens until you submit it. The software does nothing with your input on its own — it holds space for your data. Digitising paper is not AI. There is no learning, no prediction, no pattern recognition.
💳
Credit application decision
Your bank loan approved or declined in seconds
AI ?
A predictive model scored your application against patterns learned from millions of historical loan outcomes — income ratios, repayment history, location, employment type. It learned which profiles repay. This is also where data bias enters: if past lending was discriminatory, the AI learns to repeat it.
📲
Autocorrect on your phone
Your phone predicting your next word while typing
AI ?
Trained on billions of words to predict the most likely next word as you type. Modern autocorrect also learns your personal vocabulary over time. It uses the same core mechanism as ChatGPT — just at word level rather than paragraph level.
📅
Calendar reminder at 9am daily
A scheduled alarm you set yourself
Not AI ?
You wrote the rule: alert me at 9am. The system executes it exactly, every time, without variation or improvement. This is automation — it follows your instructions. AI, by contrast, figures out its own patterns from data. Knowing this difference is one of the most useful things you'll take from this course.
3
Reflect — What surprised you?
No right answers here — just your honest reaction · ~10 min

Now that you've seen the answers — what caught you off guard?

First — one quick question about your workplace:

Does your organisation currently use AI tools in day-to-day work?

Now let's go deeper

The one-sentence definition that actually works

What AI is

AI is software that learns patterns from large amounts of data and uses those patterns to make predictions, generate content, or take actions — without being explicitly programmed with rules for every situation.

That's it. No magic. No thinking. No understanding. AI is pattern recognition at massive scale. Let's unpack what that means in practice.

What "learning from data" actually means

Imagine training a new employee. You show them thousands of examples of loan applications that were approved and thousands that were rejected. Over time, they start to notice patterns: applications with certain income-to-debt ratios tend to get approved; applications from certain postal codes tend to get rejected.

AI works similarly — except instead of a person, it's a mathematical model, and instead of thousands of examples, it might be trained on millions. The model doesn't understand credit risk. It doesn't understand people. It has learned to match patterns in new applications to patterns in the training data.

The South African credit context: If historical credit data reflects the effects of apartheid — where Black South Africans were excluded from formal credit markets and forced into lower-income areas — an AI trained on that data will learn those patterns. It may then use postal code as a proxy for creditworthiness without anyone programming it to discriminate. The bias is in the data, not a deliberate choice.

The four types of AI you're most likely to encounter at work

1 Generative AI — makes new content

Examples: ChatGPT, Copilot, Gemini, Claude
What it does: Generates text, images, code, and summaries by predicting what comes next based on patterns in training data
Strength: Drafting, summarising, brainstorming, explaining
Key weakness: Cannot verify facts — it generates plausible-sounding content, not accurate content. It will confidently invent sources, statistics, and names that don't exist.

2 Predictive / Decision AI — makes classifications and scores

Examples: Credit scoring systems, fraud detection, HR shortlisting tools, spam filters
What it does: Takes inputs (a loan application, a CV, an email) and outputs a score, classification, or decision
Strength: Processing many cases at consistent speed — faster than a human reviewer for high-volume decisions
Key weakness: Reflects biases in historical training data. If past decisions were discriminatory, the model learns to repeat them.

3 Computer Vision — makes sense of images and video

Examples: Face ID, medical imaging analysis, security cameras, document scanning
What it does: Identifies objects, faces, text, or anomalies in visual data
Strength: Consistent processing of visual information at scale
Key weakness: Performance drops sharply for faces and scenarios underrepresented in training data — accuracy is often lower for darker skin tones in systems trained primarily on lighter-skinned faces.

4 Agentic AI — takes sequences of actions to achieve goals

Examples: AI assistants that research and draft reports autonomously, book travel, write and run code, send emails, or execute multi-step workflows without a human directing each individual step
What it does: Uses a language model as a “reasoning engine” to plan and carry out sequences of actions — browsing, writing, clicking, submitting — toward a stated goal
Strength: Automating complex, multi-step tasks that would normally require a human to coordinate several tools or systems over time
Key weakness: Each step is a prediction, not a guaranteed correct action — and errors compound across steps. Critically: agentic AI can take real-world actions (sending an email, modifying a file, making a booking) that are difficult or impossible to reverse. Accountability for those actions remains with the person or organisation that deployed it.

Why this matters in your workplace: As AI tools gain the ability to act on your behalf — automatically responding to emails, scheduling meetings, submitting forms — the question of “who approved this?” becomes critical. The Funda Five’s Human check is even more important when AI can act without a human in the loop.

Five things AI cannot do — regardless of how it sounds

Myth

AI knows facts — it retrieves information like a search engine

Reality

Generative AI generates plausible text based on patterns — it does not retrieve from a verified database. It cannot tell when it is wrong.

Myth

AI is neutral — it makes decisions based on pure data, without bias

Reality

AI reflects the biases in its training data. Data from unequal systems produces AI that perpetuates those inequalities.

Myth

AI understands context — it knows what you mean, not just what you said

Reality

AI processes the text you give it. It has no understanding of your organisation, your situation, or what you're trying to achieve beyond the words in your prompt.

Myth

AI can take responsibility — "the algorithm decided"

Reality

Under South African law and internationally, accountability for AI decisions rests with the people and organisations that deploy them — not the software.

Myth

Confident output = correct output

Reality

AI outputs are expressed with the same tone and style whether they are accurate or completely fabricated. Confidence is not a signal of accuracy.

Knowledge Check 5 questions · check your understanding

You need to answer all 5 questions correctly to move on. You can retry as many times as you like.

1. What best describes how AI works?

2. A bank's AI keeps declining applications from certain postal codes. What's the most likely cause?

3. Which type of AI would most likely be used to automatically shortlist job applications?

4. A generative AI tool gives you a detailed, confident response. What should you do?

5. When an AI system makes a harmful decision, who is legally responsible?

4
Apply — AI in your workplace
Take it back to your own job · ~10 min

You've seen what AI is, what it isn't, and the three types you're most likely to encounter. Now bring it to your world.

Answer these three questions about your own workplace: