AI Basics Lab 🤖

General
1

Welcome :waving_hand:

Artificial Intelligence (AI) is already part of everyday life, whether we realise it or not. It shows up in our phones, our social media feeds, our workplaces, our schools, our banks, our customer service experiences, and the digital systems around us.

This lab is here to break AI down in a way that is simple, practical, and connected to real life. The goal is not just to learn what AI is, but to build confidence, think critically, and understand how we can engage with it in a more informed and grounded way.

For Pacific communities, this matters deeply. AI is not just another tech trend. It is becoming part of how information is shared, how services are delivered, how stories are told, and how people and cultures are represented online.

If we do not understand AI, we risk being shaped by systems we did not build and may not fully understand. This space is about making sure our people are not left behind, misrepresented, or overlooked in the age of AI.

This is not about fear.

It is about awareness.

It is about confidence.

It is about learning enough to ask better questions.

What you’ll learn in this lab :open_book:

  • what AI is

  • how AI works

  • how AI reads patterns

  • the difference between machine learning, deep learning, generative AI, and Large Language Models

  • where AI is already showing up in everyday life

  • how Pacific communities can use AI with confidence and care

Why this lab matters

A lot of AI content online focuses only on tools.

It says:

  • here is ChatGPT

  • here is a prompt

  • here is how to save time

But that is only the surface. And tools change all the time.

This lab goes deeper. Because once you understand the engine underneath, AI becomes less mysterious. And once it becomes less mysterious, it becomes easier to question.

What is AI? :robot:

AI stands for Artificial Intelligence.

In simple terms, AI is when machines or computer systems are designed to do tasks that normally require human intelligence.

This can include:

  • recognising patterns

  • answering questions

  • making recommendations

  • translating language

  • generating writing or images

  • summarising information

  • predicting outcomes

  • helping automate tasks

AI learns from data. It looks for patterns and uses those patterns to make a prediction, recommendation, or output.

AI is not magic.

It is not spirit.

It is not wisdom.

It is not lived experience.

It is not culture.

It is a tool.

And like any tool, it depends on:

  • what it was trained on

  • what data it learned from

  • how it was built

  • how people use it

  • whether the output is checked carefully

At its core, AI is about patterns :dizzy:

One of the most important things to understand about AI is this:

AI works by reading patterns.

It looks at lots of examples and starts noticing what usually happens, what often comes next, what seems connected, and what looks unusual.

For example:

  • Maps reads traffic patterns and suggests the fastest route

  • Spotify notices listening patterns and recommends songs

  • banks notice spending patterns and flag suspicious activity

  • writing tools notice language patterns and predict likely words

So when we say AI is “intelligent,” a lot of what that really means is that it is very good at noticing patterns in large amounts of data.

That is powerful. But it is also important to keep it in perspective.

AI is often not “thinking” in the human sense. It is reading patterns, making predictions, and responding based on probability.

Patterns are not new to our people :ocean:

For Pacific people, pattern recognition is not new.

Our ancestors were already masters of reading patterns in the natural world. They observed the tides, currents, winds, stars, clouds, seasons, bird movements, and the behaviour of fish. Through deep observation over generations, they built knowledge of how land and sea work together.

In fishing, for example, our ancestors did not just go out blindly and hope for the best. They knew how to read the environment. They understood that changes in moon phases, tides, winds, currents, bird behaviour, water conditions, and seasonal rhythms could all signal where fish may be, when the right time was to go out, and when it was better not to.

That is pattern recognition.

They were learning from repeated signals in the environment:

  • when this happens, that often follows

  • when the sea moves this way, conditions are changing

  • when birds gather in certain places, fish may be near

  • when the moon and tides align in certain ways, it affects the sea

  • when the land, sky, and ocean behave in certain patterns, they tell a story

That knowledge did not come from guessing. It came from observation, memory, repetition, experience, and relationship.

In a way, AI works similarly in structure, but not in spirit. AI also looks at patterns and uses them to make predictions.

AI reads patterns in data.

Our ancestors read patterns in life.

Our ancestors carried context, wisdom, memory, place, responsibility, and relationship. Their pattern recognition was grounded in care for people, community, and survival.

That is why Pacific knowledge still matters in the age of AI.

A simple way to remember it

  • AI = patterns
  • Machine Learning = learning from examples
  • Deep Learning = learning more complex patterns through layers
  • Generative AI = creating new content from learned patterns
  • Large Language Models = generating language from learned patterns

From a Pacific lens:

Our ancestors read patterns in the land, sea, winds, stars, and seasons.

AI reads patterns in data.

Both involve patterns. But only one carries ancestral memory, lived experience, cultural meaning, and responsibility.

How AI works in simple terms :melting_face:

1. Data goes in

AI systems are trained on data.

This data can include:

  • words

  • images

  • numbers

  • transactions

  • videos

  • speech

  • behaviour patterns

  • other kinds of information

2. The system looks for patterns

The model starts noticing repeated relationships and patterns in that data.

3. It learns what is likely

Over time, it gets better at predicting what usually comes next, what matches, what seems unusual, or what output fits a request.

Examples:

  • “Peanut butter and…” → “jelly”

  • “Happy birthday…” → “to you”

4. It produces an output

That output could be:

  • a recommendation

  • a prediction

  • a summary

  • a written response

  • an image

  • an alert

  • a suggested action

5. Humans still need to interpret it

This part matters most.

The output still needs:

  • judgement

  • review

  • context

  • accountability

  • and sometimes cultural understanding that AI simply does not have

Breaking it down :axe:

Machine Learning :gear:

Machine Learning is when computers learn from data instead of being programmed with every single rule.

Instead of saying,

“if this happens, then do exactly this,”

you give the system lots of examples, and it learns the pattern.

Real-life example:

A bank trains a system on lots of transaction data so it can learn what normal spending looks like and what might look suspicious.

Simple way to think about it:

Learning by example.

Deep Learning :brain:

Deep Learning is a more advanced type of machine learning that uses multiple layers to recognise more complex patterns.

It is often used for things like:

  • image recognition

  • facial recognition

  • speech recognition

  • language understanding

  • complex predictions

Real-life example:

Your phone recognising your face to unlock is often powered by deep learning.

Simple way to think about it:

A more layered and complex form of machine learning.

Generative AI :artist:

Generative AI is a type of AI that creates new content based on patterns it has learned.

That can include:

  • writing

  • images

  • music

  • code

  • summaries

  • ideas

  • audio

  • video

Real-life example:

When you ask ChatGPT to draft an email, explain a concept, or create a workshop outline, it is generating something new based on language patterns it has learned.

Simple way to think about it:

Using learned patterns to create new content.

But remember: it is not creating from human wisdom. It is creating from pattern-based prediction.

What are LLMs/Large Language Models? :robot:

Large Language Models, often shortened to LLMs, are a type of AI trained on huge amounts of text so they can understand and generate human-like language.

That means they are trained to work with:

  • words

  • sentences

  • conversations

  • articles

  • books

  • websites

  • questions and answers

  • patterns in language

Tools like ChatGPT, Claude, and Gemini are powered by Large Language Models.

In simple terms, an LLM is a system that has learned from massive amounts of language data and uses those patterns to predict what words are most likely to come next.

That is why they can:

  • answer questions

  • explain ideas

  • write emails

  • summarise information

  • brainstorm content

  • help with planning

  • generate stories or scripts

Why are they called “large”?

They are called large because they are trained on very large datasets and have a huge number of parameters.

The easiest way to think about it is:

A Large Language Model is a very big language pattern reader.

How do LLMs work? :nerd_face:

They:

  1. learn from massive amounts of text

  2. notice patterns in how language is used

  3. predict what word, phrase, or idea is likely to come next

  4. generate a response based on those probabilities

So when you ask a question, the model is not sitting there thinking like a human. It is generating the most likely useful response based on the patterns it has learned.

Simple example

If you type:

“The sun rises in the…”

a language model can predict that the next word is likely “east.”

That same kind of prediction happens at a much bigger scale across billions of words and examples.

Why LLMs feel so smart :nerd_face:

They are:

  • fast

  • fluent

  • good at sounding natural

  • able to pull together patterns from lots of text

But they can also:

  • make things up

  • get facts wrong

  • miss context

  • flatten culture

  • sound confident while still being inaccurate

Why this matters for Pacific communities :hibiscus:

Large Language Models are often trained on broad internet data and large bodies of text that may not fully represent Pacific languages, cultures, histories, or ways of knowing.

That means they can:

  • confuse Pacific cultures with one another

  • mispronounce or misspell names

  • flatten deeper meanings

  • provide generic answers to cultural questions

  • sound respectful while still being wrong

Just because an LLM can generate language does not mean it carries cultural understanding.

It can be useful. It can be powerful. But it still needs human judgement, cultural knowledge, and care.

What AI is not

AI is not:

  • a human being

  • a cultural expert

  • automatically truthful

  • always fair

  • free from bias

  • something you should trust blindly

AI can sound polished, smart, and confident while still being wrong.

That is why human judgement still matters.

That is why community knowledge still matters.

That is why elders, teachers, knowledge holders, and lived experience still matter.

AI in everyday life :people_holding_hands:

You may already be using AI when you:

  • use Google Maps for traffic and route suggestions

  • get music recommendations on Spotify

  • get video recommendations on YouTube or Netflix

  • use predictive text or spell check on your phone

  • unlock your phone with face recognition

  • receive fraud alerts from your bank

  • talk to a chatbot on a website

  • scroll social media feeds shaped by algorithms

  • use AI tools like ChatGPT, Gemini, Claude, or Perplexity

So even if someone says, “I don’t use AI,” they often already do.

Real-life examples

Google Maps

Google Maps reads traffic patterns, road conditions, and travel behaviour to suggest the fastest route.

Helpful because it saves time.

But it still does not know everything. Local knowledge still matters.

Spotify or YouTube

These platforms watch your behaviour and suggest content based on what you seem to like.

Helpful because it feels personalised.

But it can also narrow what you see and keep you in a certain content bubble.

Bank fraud detection

Banks use AI to learn what normal spending patterns look like and what seems unusual.

Helpful because it can protect people from fraud.

But systems can still make mistakes and flag normal activity.

Predictive text

Your phone learns common language patterns and starts suggesting your next word.

Helpful because it saves time.

But it can also change meaning or make odd mistakes.

Chatbots

A website chatbot answers questions by drawing from patterns in information and past interactions.

Helpful because it can respond quickly.

But it may not understand complexity, emotion, or nuance.

Pacific and community examples

Community event planning

A Pacific community group can use AI to draft a flyer, social media caption, event outline, or funding blurb for a youth event, church programme, cultural performance, or community workshop.

Helpful because many community groups are time-poor and under-resourced.

But the final message still needs to sound like the real voice of the community, not generic internet language.

Pacific language and cultural meaning

Someone asks AI to explain a Pacific word, value, saying, or cultural practice.

It may give a quick answer. But it can also:

  • confuse one culture with another

  • flatten deeper meaning

  • strip away context

  • sound respectful while still being wrong

This is why knowledge holders still matter.

Pacific youth using AI for school

A young person might use AI to explain homework, generate study questions, practise writing, or structure a speech.

Helpful because it can support confidence and learning.

But they still need to:

  • fact-check

  • avoid copying blindly

  • keep their own voice

  • learn when AI is wrong

Small Pacific business

A Pacific-owned food business, clothing brand, or side hustle could use AI to:

  • write Instagram captions

  • draft menu descriptions

  • generate customer replies

  • brainstorm ideas

  • organise basic admin

Helpful because it saves time and supports small operators.

But the tone, values, and identity of the business still need to come from the people behind it.

Community services

Pacific families may interact with AI through:

  • online forms

  • customer support bots

  • digital appointment systems

  • healthcare portals

  • government service systems

Helpful because these systems can improve speed and access.

But they can still fail if they do not understand real-life complexity, language needs, or culturally grounded communication styles.

Pacific storytelling and culture

AI can help brainstorm workshop ideas, lesson content, scripts, digital resources, or community storytelling concepts.

Helpful because it supports creativity and planning.

But culture is not just content. Some knowledge needs:

  • context

  • care

  • consent

  • cultural guardianship

  • human responsibility

Not everything should be passed through AI just because it can be.

What is Generative AI useful for?

Generative AI can help with simple everyday tasks like:

  • drafting emails

  • rewriting messages

  • summarising long text

  • brainstorming ideas

  • planning a workshop

  • writing captions

  • creating checklists

  • helping explain hard concepts

  • turning rough notes into something clearer

That is why many people are drawn to it. It can be practical, accessible, and fast.

But fast does not always mean right.

Benefits of AI

When used well, AI can:

  • save time

  • support learning

  • help organise ideas

  • assist with creativity

  • improve productivity

  • reduce repetitive admin

  • help small teams do more with fewer resources

  • make information easier to access

For communities with limited time, money, and support, these benefits can be meaningful.

Risks to be aware of:

Misinformation
AI can produce answers that sound right but are false.

Bias
AI can reflect unfair patterns from the data it was trained on.

Cultural misrepresentation
AI may mix up Pacific cultures, oversimplify meanings, or get names, histories, and contexts wrong.

Privacy
People may upload personal or confidential information without realising the risk.

Overreliance
People may begin trusting AI more than their own thinking or the wisdom of others.

Loss of voice
If people depend too much on AI, their own voice, style, and creativity can weaken.

Unequal impact
Not every community benefits equally from AI. Some communities may face more exclusion, misunderstanding, or harm than others.

How to use AI well

  • treat AI as a support tool, not the final authority

  • double-check important facts

  • do not upload private, personal, or confidential information

  • be careful with cultural knowledge and context

  • use your own voice and judgement

  • question answers that seem too polished or too certain

  • ask what kind of AI you are using: predictive, analytical, or generative

  • think about what patterns it may be relying on

  • remember that human context still matters

Reflection questions

  • Where am I already seeing AI in daily life?

  • What patterns might that system be reading?

  • What kind of data is it learning from?

  • What do I find useful about AI?

  • What worries me about it?

  • Where could AI get Pacific people or Pacific knowledge wrong?

  • What kind of knowledge should never be handed over without care?

  • What should always stay with human judgement, community accountability, and cultural understanding?

Simple activity: Spot the pattern

Think about one everyday AI example, like Maps, Spotify, TikTok, or ChatGPT.

Then ask:

  • What patterns is this system reading?

  • What data is it relying on?

  • Is it helping me, influencing me, or both?

  • What could it get wrong?

  • What would human judgement add here?

You can also compare it to ancestral knowledge:

  • What signs did our ancestors watch in the environment?

  • What did those patterns help them decide?

  • What is the difference between reading data and reading life?

That is where deeper learning begins.

Key takeaway

AI is powerful, but it is not magic.

At its core, AI works by reading patterns, learning from data, and generating outputs based on probability.

For Pacific people, that idea should not feel completely foreign. Our ancestors already knew how to observe patterns deeply through the land, sea, winds, stars, and seasons. They understood that patterns carry meaning, but only when interpreted with context, experience, and responsibility.

AI can read patterns.

But people still carry wisdom.

And that wisdom matters more than ever.

Coming next

Next, we move into Data Governance Foundations.

Because if AI learns from patterns in data, then the next big questions are:

  • where did that data come from?

  • who owns it?

  • who is responsible for it?

  • how is it managed?

  • how is it protected?

  • what happens when it is wrong?

AI is only as strong, fair, and trustworthy as the data underneath it.

For Pacific communities, that makes data governance essential.