AI Around Us

Artificial Intelligence is already part of your morning routine — before you even think about it. You unlock your phone with your face, your music app opens with a playlist it picked just for you, and Google Maps quietly reroutes around a traffic jam on Sukhumvit. Minutes later, GrabFood suggests your usual pad kaprao, LINE offers to translate a message from a friend overseas, and Facebook shows you the one post your best friend will laugh at. None of this is magic — it’s AI, running invisibly in apps you already trust. In this mission, you’ll learn to spot where AI is hiding and understand what it’s actually doing behind the scenes.

When people say “AI”, they often picture a robot from a movie — a mind in a metal body. The reality is much less dramatic, and much more useful. The field of AI has been around since the 1950s, when researchers first asked whether machines could be made to think. For decades, progress was slow: early AI ran on hand-written rules — if-this-then-that instructions that quickly broke when the real world didn’t cooperate.

The real shift came from a different idea. Instead of writing rules, what if we let the computer learn patterns from examples? This approach — called machine learning — only became practical once three things arrived together: enough data (from phones and the internet), enough computing power (cheap GPUs), and better algorithms. Around 2012, those pieces lined up — notably when deep learning breakthroughs demonstrated the power of this approach — and AI took off.

That’s why AI today looks very different from the clunky chatbots of the past. Your Shopee “You may also like” feed, the translated captions on YouTube, even the fraud alert from your bank app — all of them work because a model has seen millions of similar examples and learned to spot the pattern. It’s not thinking like a human. It’s just very, very good at finding patterns in data.

Recommendation systems predict what you’ll probably like next. They compare your past behavior with millions of other users to spot patterns. When Shopee lines up “recommended for you” items after you view one product, or Netflix pushes a series to your home screen, a recommendation system is at work. The more you interact, the more it learns what you actually click on versus what you scroll past.

Computer vision lets machines interpret images. Your phone’s face unlock, your banking app’s biometric login, and Face ID on an iPhone all use a camera plus a model trained on millions of faces. Beyond phones, computer vision powers Kerry and Flash scanners reading a parcel label in a fraction of a second, and medical imaging at hospitals that helps doctors spot abnormalities.

Natural Language Processing (NLP) is how machines understand and produce human language. The translate button inside LINE, Siri taking a voice command, and ChatGPT writing a whole email all sit on top of NLP models. Recent large language models have made this feel nearly conversational — but underneath, the model is still predicting which word most likely comes next, not actually knowing what it’s saying.

Classification is the quiet workhorse. It puts each new input into a category it has learned. Gmail’s spam filter classifies every incoming email as spam or not. The alert when your bank app flags a suspicious transaction is a classification model deciding “likely fraud” versus “normal”. Shopee categorizes seller listings, Facebook labels potentially harmful content — same idea, different data.

Four ideas, one pattern: each system takes a specific kind of input, learns from examples, and produces a focused output. Once you know the four, you can usually guess which one is running inside any app you open.

Notice how the matching game you just played was really a test of this same idea: each service on the left is a specific job, each AI technology on the right is a specific method. The art isn’t knowing every product on the market — it’s recognizing which of the four patterns fits the problem at hand. That’s a skill that holds up even as new apps and new AI features launch every week.

Recognizing AI when you see it isn’t just a party trick — it’s the foundation for everything else. Once you know GrabFood, Google Maps, and your bank app are all running narrow AI models, you can start asking better questions: What data is this using? What might it get wrong? Whose choices shape what it shows me?

AI touches parts of daily life that feel purely administrative. Digital wallet apps commonly use pattern detection to flag unusual transactions, and health systems use classification models to prioritize alerts. These aren’t futuristic — the same families of models you just met are running today, on services you probably already use.

In the next mission, “How Does AI Work?”, you’ll go one layer deeper. Instead of just spotting AI, you’ll see how a model actually learns — what training data means, how labels shape behavior, and why bad data produces bad AI no matter how clever the algorithm is. The stuff you noticed today is built from exactly those pieces.