Anu's AI Spy Day 🕵️
After her first lesson, Anu told her friend Priya about AI. Priya wasn't convinced. "It's just apps," she said. "Nothing special."
Anu decided to prove it. She kept a small notebook with her for one full day and wrote down every time she thought AI was involved. By lunch she had 6 things. By bedtime she had 14.
Priya read the list and went quiet. "Wait — even the weather app?" she asked.
"Especially the weather app," said Anu. "And face unlock. And every time YouTube suggested a video. And autocorrect. And Maps knowing about the traffic even before Appa checked. And—"
"Okay, okay," said Priya. "Teach me."
📺 YouTube — The Recommendation Machine
When you open YouTube, the first thing you see is not a random list. Every single video on that page was chosen specifically for you by an AI system that has been watching what you do.
Here is what the YouTube AI notices about you:
Watch Time
How many seconds of each video did you actually watch? Did you rewind parts? Or did you skip ahead or stop early?
Completion Rate
Did you watch a video all the way to the end? A video watched fully tells the AI you loved it — even if you didn't click Like.
What You Click Next
After watching, did you click on a suggested video or go back? What you click next tells the AI what you want more of.
Similar Viewers
The AI groups you with millions of people who watched the same things. Whatever those similar people loved next — it suggests that to you too.
Smart way to use YouTube
- Search for what you actually need
- Use it for learning with intention
- Set a timer before you start
- Notice when you have gone off-topic
What the rabbit hole looks like
- Opening YouTube "just for 5 minutes"
- Following "Up Next" without thinking
- Watching 11 videos when you planned 1
- Forgetting what you originally came for
🔍 Google Search — More Than Just Keywords
When you type something in Google and press Enter, you are not just looking up words in a dictionary. You are using one of the most complex AI systems ever built.
Two big things Google AI does that most people don't know:
- It reads the meaning, not just the words. If you type "best food for exam day", Google understands you want healthy, brain-boosting food ideas — not a list of every page that contains those exact words. This is called semantic search.
- It personalises your results. Two students can search the exact same thing and see different results — because Google remembers their past searches, location, what device they use, and even what language they usually read in.
Google also now shows AI-generated answers at the very top of results (called AI Overviews). These are written by AI in real time. They can be very helpful — but they can also be wrong. Always check the sources linked below the AI answer.
🗺️ Google Maps — The Traffic Predictor
When your family checks Google Maps before leaving home, and it says "there's heavy traffic on NH-65 — try Ring Road instead" — that is AI at work.
How does Maps know about traffic that is happening right now?
Crowdsourced Speed Data
Every phone with Google Maps open (and location enabled) sends its GPS speed to Google. Millions of phones on every road = real-time traffic map.
Historical Patterns
Maps remembers: every Monday morning at 8:30 AM, the road near this school is jammed. This historical pattern gets added to the prediction.
Local Events
If a cricket match or festival is happening nearby, Maps knows and adjusts. It learns from past events to predict future congestion.
Real-Time Incidents
Users can report accidents. AI analyses patterns in sudden slowdowns to detect unreported accidents and adjust routes automatically.
This is why Maps can say "arrive by 9:15 AM" with surprising accuracy. It calculated that ETA in milliseconds by running thousands of possible routes through AI.
🔓 Face Unlock — AI Recognising You
When you pick up a phone and it unlocks just by looking at your face — that is biometric AI. "Biometric" means using your physical features (face, fingerprint, voice) as a password.
🪄 How does face unlock work? Three steps:
- Setup: The phone takes many photos of your face from different angles.
- Mapping: AI builds a mathematical map — the gap between your eyes, the shape of your nose, the curve of your jaw.
- Matching: Each time you pick up the phone, this map is compared against your face in less than half a second.
- Bright sunlight or very dark rooms confuse the camera
- If you are unwell and your face looks different
- Early face unlock systems could be fooled by a photo — modern ones use infrared sensors to check for depth (so a flat photo does not work)
🇮🇳 Other biometric AI in India:
- Aadhaar eKYC — face matching to verify your identity
- Bank phone lines — voice recognition to confirm you are the account holder
- Some railway stations — face recognition at entry gates
⌨️ Autocorrect and Predictive Text — Your Smart Keyboard
Every time you type on your phone's keyboard and it fixes a spelling mistake or suggests the next word — that is a small but powerful AI model running on your device.
How does predictive text work? The keyboard AI has studied billions of real sentences in English (and in your language, if enabled). It learned which words usually come after which other words.
For example: after "I am going to the" — the most likely next words are "school", "market", "hospital", or "station". So those are suggested.
When autocorrect helps
- Fixing typing speed errors
- Learning new vocabulary spelling
- Finishing common phrases faster
- Swipe typing (slide finger over letters)
When autocorrect causes trouble
- Changing a correct word to a wrong one
- Changing names it doesn't know
- Making embarrassing changes in messages
- Fixing technical terms incorrectly
🌦️ Weather Apps — AI that Reads the Sky
Your phone's weather app knows it will rain at 3 PM today in your specific area. Twenty years ago, this level of precision was impossible. Today, AI-powered weather forecasting has changed everything.
What data does weather AI use?
- Satellites — India's INSAT and IRNSS satellites send images of cloud movements and temperatures every 15 minutes
- Weather stations — Thousands of sensors across India measure temperature, humidity, wind speed, and air pressure on the ground
- Ocean buoys — Sensors in the Bay of Bengal and Arabian Sea detect early signs of cyclone formation
- Historical records — 100+ years of India weather data teaches the AI what patterns repeat every monsoon season
- Radar — Doppler radar scans rain clouds and measures how fast they are moving toward which locations
The AI does not "know" the weather the way a person knows things. It runs mathematical models: "given all the data inputs right now, what is the most probable atmospheric state in 6 hours, 12 hours, 24 hours?" It runs millions of calculations per second to answer that question.
🫧 The Recommendation Trap — Filter Bubbles
Here is a problem that no one talks about enough — and it is caused by recommendation AI working perfectly.
A filter bubble happens when an AI keeps showing you content that matches what you already believe, like, and agree with — until you stop seeing different ideas at all.
Example: Ravi watches one video supporting one political party. YouTube shows him another. Then another. After a month, his feed shows only those views. He starts to think "everyone thinks this way" — but in reality, he has been put inside a bubble. The AI never showed him the other side, because the algorithm found he did not click on opposing views.
How to break out of a filter bubble:
- Deliberately search for views you have not heard before
- Follow channels or accounts with different perspectives
- Notice when you are only seeing one side of any debate
- Ask: "What would someone who disagrees say about this?"
- Read actual news articles rather than social media summaries
🧠 You Are Now an AI Spotter — 5 Habits of Smart Users
By the end of this lesson you know that AI is everywhere around you — in apps, maps, keyboards, cameras, weather, and your social feeds. Now the question is: what do you do with that knowledge?
Spot it and name it
When you use any app, ask yourself: "Is AI involved here?" Over time this becomes automatic — and you stop being surprised or fooled by AI.
Search with intention
Don't let the recommendation engine drive your time. Decide what you want to watch or read before you open YouTube or a news app. Search directly.
Verify before you trust
AI Overviews, weather forecasts, and WhatsApp forwards can all be wrong. Always check the original source before trusting important information.
Check your privacy settings
Know which apps have access to your camera, location, and contacts. Ask your parent to review these with you once every few months.
Seek different viewpoints
Deliberately consume content from different sources on important topics. Break your own filter bubble before it forms. This makes you a stronger thinker.
🎯 Quick Quiz — 10 Questions to Check What You Learned
📝 Activity — Anu's AI Spy Day Challenge
Tip: in the print dialog, choose “Save as PDF” to download instead.Just like Anu did in the story, spend one full day noticing AI around you. Use this table to record what you find. Write in your notebook or print this page.
| # | App or Tool | What did the AI do? | Type of AI | Was it helpful? Y / N |
|---|---|---|---|---|
| 1 | ||||
| 2 | ||||
| 3 | ||||
| 4 | ||||
| 5 | ||||
| 6 | ||||
| 7 | ||||
| 8 |
Reflection questions (write in your notebook):
- Which AI tool surprised you most? Why?
- Did you notice a filter bubble at work during the day? Describe it.
- Name one AI tool you will use more intentionally after this lesson.
Use this table in your notebook today, or print this page directly if helpful.
What this lesson covers: This is Lesson 2 of 12 in the Class 6 full-year AI curriculum. Students explore AI in the apps they use every day — YouTube recommendations, personalised Google Search, Google Maps traffic prediction, face unlock, autocorrect / predictive text, and weather forecasting. The lesson introduces filter bubbles and builds five smart-user habits.
Learning time: Around 60–75 minutes, or split across two sessions. The "AI Spy Day" worksheet activity works well as a one-day homework assignment.
Safety by design:
- No personal data collected from students on this page.
- No login or sign-up required.
- Privacy and safety points are woven into every section, not just mentioned once.
- The filter bubble section builds media literacy — a core safety skill for digital citizens.
Great conversation starter: Ask your child tonight: "Can you find 3 AIs you used today without realising it?" Then listen — you may be surprised by what they know now.
For classroom use: The AI Spy Day activity works brilliantly as a 2-day project: Day 1 introduce the concepts, Day 2 students share their spy lists. Class discussion on filter bubbles can be a powerful critical thinking exercise.