You are already a daily AI user
Before you ever opened this course, before you heard the term machine learning, before anyone explained what a neural network was, you were already interacting with AI dozens of times a day. The question was never whether AI would enter your life. It already had. The question was whether you understood what was happening.
This lesson is about making the invisible visible. We are going to walk through a single ordinary day and identify every moment where AI is making a decision, forming a prediction or shaping your experience. By the end, you will not be able to look at your phone the same way again.
Why this matters for learning
One of the biggest barriers to learning AI is the feeling that it is happening somewhere else, in research labs or tech companies far removed from ordinary life. Once you see that AI is already embedded in your morning routine, your commute and your entertainment, that barrier dissolves.
Everything you learn in this course is already operating in the world. You are not studying something abstract. You are learning to understand something you already live inside.
A day in your life, with the AI visible
Let us trace through an ordinary day. Not a dramatic one. A Tuesday. The kind where nothing particularly exciting happens. Watch how many times AI appears.
7:00AM
Your alarm goes off. You pick up your phone.
Face ID or fingerprint recognition unlocks it instantly. A deep learning model trained on thousands of facial images or fingerprint patterns is comparing what it sees to what it learned about your specific face or thumb. If the light is different, if you are squinting, it still works. That is the neural network adapting, not a fixed rule being checked.
Computer vision
7:10AM
You check your email.
Before you see a single message, a machine learning model has already scanned every incoming email and decided which ones are spam, which are promotional and which are important. It learned those categories from billions of labelled emails and from your own behaviour over time. When you mark something as spam, you are training it.
Spam filtering
7:30AM
You open Spotify or Apple Music on the way to the kitchen.
The playlist that autoplays was assembled by a recommendation system that analysed your listening history, the listening history of people with similar taste, the audio features of songs you have liked and the time of day. It made a prediction about your mood at 7:30 on a Tuesday morning without you asking it to.
Recommendation
8:15AM
You check Google Maps before leaving.
The estimated arrival time is a prediction generated by a model trained on years of traffic patterns, real-time data from millions of other phones on the road, historical patterns for this specific road at this specific time on this day of the week. It is not looking at a map. It is making a statistical prediction.
Prediction
12:30PM
You pay for lunch with your card.
At the moment your card is processed, a fraud detection model analyses this transaction against thousands of data points: your usual spending patterns, the location, the merchant type, the time of day, the amount. It makes a decision in milliseconds. If anything looks unusual, it flags or blocks the payment before you even get a receipt.
Fraud detection
1:00PM
You do a Google search.
The ten results you see were ranked by an algorithm that weighs hundreds of signals: the relevance of the page to your query, the authority of the source, your location, your search history and what people with similar queries clicked on. Natural language processing interprets your words, even if they are vague, misspelled or phrased as a question.
Search ranking
7:00PM
You open Netflix.
The thumbnails of shows change depending on who is watching. Netflix tests different images for the same title and shows you the version their model predicts you are most likely to click. The entire homepage is a personalised grid assembled in real time, not a fixed catalogue. Your row of recommendations is different from your neighbour's.
Personalisation
9:30PM
You send a voice message on WhatsApp or use Siri.
Your voice is converted to text by a speech recognition model, a deep learning system trained on thousands of hours of human speech across accents, ages and recording conditions. The model does not listen for specific words. It predicts the most likely sequence of words given the sounds it heard.
Speech recognition
The six categories of everyday AI
Looking across that day, the same types of AI keep appearing under different names and in different products. Here are the six categories that power most of the AI you interact with daily.
Predict what you will want to see, listen to or buy next based on your history and the behaviour of similar users.
Netflix · Spotify · YouTube · Amazon · TikTok · Instagram
Enables machines to understand and interpret images and video. Includes facial recognition, object detection and medical imaging.
Face ID · Google Photos · Snapchat filters · Tesla Autopilot
Natural language processing
Helps machines read, understand and generate human language. Powers search, translation, autocomplete and AI assistants.
Google Search · Translate · Grammarly · Siri · ChatGPT · Claude
Use historical patterns to forecast outcomes: traffic, demand, weather, health risks, financial behaviour and much more.
Google Maps ETA · Weather apps · Fraud detection · Credit scoring
Automatically classify and sort content at scale. Makes decisions about what you see and what gets removed without human review.
Spam filters · Content moderation · Ad targeting · Email sorting
Creates new content: text, images, audio, code and video. The fastest-growing category in AI and the one most visible to the public right now.
ChatGPT · Claude · DALL-E · Midjourney · GitHub Copilot · Suno
The things happening without your permission or awareness
Most of the AI in your life operates entirely in the background. You do not opt in. You do not see the decision being made. You just experience the output, a filtered inbox, a personalised feed, a blocked transaction, a suggested route.
This is worth pausing on, not to alarm you, but because it is the context for everything that follows in this course. When you understand how these systems work, you become a participant rather than just a subject. You can ask better questions, make more informed choices and eventually build systems that treat the people using them with the same care you would want for yourself.
01
The order of your social media feed
No social media feed is chronological by default anymore. Every post you see has been ranked by a model predicting which content will keep you engaged longest. The ranking changes every time you open the app.
02
The price you see for flights and hotels
Dynamic pricing models adjust what you are shown based on your device, your search history, the time of day, demand patterns and how urgently the algorithm thinks you need to book. Two people searching the same route can see different prices.
03
Whether your job application gets a human review
Many large companies use AI to screen CVs before a recruiter ever reads them. The model scores applications against patterns in past successful hires. If your CV does not match those patterns, it may never reach a human being.
04
The autocomplete suggestions in your keyboard
Your phone keyboard is running a small language model that has learned your writing patterns and predicts the next word you are likely to type. It personalises over time based on your actual messages. It is a language model, just a much smaller one than GPT or Claude.
05
Your credit limit and loan interest rate
Financial institutions use machine learning models to assess creditworthiness. The model looks at hundreds of variables beyond just your income and credit score. The interest rate you are offered may have been set by a model with no human involved in the final decision.
What changes now that you know this
Awareness is the first form of power. Once you can see where AI is operating, you can start asking the right questions: who built this system, what data did it learn from, what is it optimising for, and whose interests does it serve?
A spam filter optimising to remove unwanted email serves you. An engagement algorithm optimising to keep you on a platform as long as possible may serve the platform more than it serves you. Both are AI. Neither is inherently good or bad. The difference lies in what they were designed to do and who benefits.
As you build your own AI systems throughout this course, you will make decisions like these. What should your model optimise for? What counts as success? Who is affected by the output? These questions begin here, with the realisation that AI is not neutral. It always reflects the choices of whoever built it.
Phase 1 complete
You now have the full picture.
You understand where AI came from, how the three terms relate to each other, what types of AI actually exist today and where AI already lives in your daily life. Phase 2 is where we get into the building blocks: data and mathematics, the raw material that makes AI possible.