| Computational Thinking |
| DecompositionPattern RecognitionAlgorithmic Thinking | - Jungle fruit party: each guest eats only fruits they like and needs three to feel full
- Draw a house where roof, windows and door are each a different shape
| - A matchstick equation that doesn't balance: which sticks to remove
- Fill the missing steps when only the start and end are shown
| - Invented chatbot words (Crumblerific, Plantification): which one has no real root
- Decode by splitting a compound word or sequence into parts
| - Work backwards from a known end state to the start
- Handle nested, multi-rule structures
|
| Pattern RecognitionPattern Recognition | - A robot sorts images into Stripes, Dots or Everything-else: which is misplaced
- A shape sequence rotates and grows: what comes next
| - Find the rule from before-and-after pairs, apply it to a new shape
- Place objects in a 2D table by two properties
| - A cipher shifts each letter by its position (CHILD becomes DJLPI): encode RIVER
- AI tokenisation: which way the words split
| - Learn a notation taught in the question; read a diagram into notation and back
|
| Abstraction (CBSE frames it spatially)Spatial ReasoningPattern RecognitionG5 up, some papers; via Pattern Recognition at G3-4 | - Mentally rotate and flip a shape to match the pattern
- Fit the pieces into a grid
| - Fold, cut and unfold paper: predict the result
- Mirror reflections
| - Stacked shapes seen from the top: order them bottom to top
- Rotate cards to bring all the stars to the centre
| - Multi-step 3D transformations with measurement constraints
|
| Algorithmic ThinkingAlgorithmic Thinking | - Relay action game: each player repeats the last action, who got it wrong
- Popcorn 4 min, lemonade 2 min, started together: when are both ready
| - Order matters: A then B vs B then A; predict a loop's result
- Block code: which blocks draw the given pattern
| - Conditional golf: IF moon go right, ELSE go up, which layouts reach the goal
- Move, Fill-dot and Repeat blocks: what grid is drawn
| - Trace pseudocode with variables and arithmetic
- Reverse a procedure to recover the start state
|
| Artificial Intelligence CBSE from Class 6; ASSET assesses AI reasoning from Grade 3 |
| What AI is and how it worksAI Foundations / Technologies | - A device's parts decide what it can do (mic hears, camera sees)
- Images shown as number grids; brightness adds to each value
- A rule-based sorter classifies a new object, even when 'wrong' in real life
| - Two chatbots guess a missing word: reason how each one predicts
- Detection (is something there) vs identification (what is it)
| - Which tasks need a learning AI versus a fixed rule
- Number representations, and how a swapped value changes the output
| - An AI 'temperature' setting trades predictability for creativity: pick the right one
- Compare two ranking algorithms on the same data
|
| AI as algorithm / neural networksAI TechnologiesCBSE Class 8 | | | - Neural-network nodes each detect a digit part: recognise a 9 when exactly two fire
| - Extended node sets: reason how layered detectors combine into a decision
|
| AI methodologies, how AI learnsAI Training & Learning | | - Self-learning maze robot, +1 for the exit and -1 for a wall: what it does in a new maze
- Which features help a model learn versus add noise
| - An AI draws watches almost always at 10:10: best explanation is the training data
- Predict what a model trained on one era gets right and wrong
| - A model trained to Sep 2021: which March 2022 questions it can answer
- How a recommender's criteria change over time
|
| Data for AIData AnalysisAI Training & Learning | - Read a table, cross-reference rows and columns, apply a condition
| - Test values against two numeric conditions at once
- Tell a warranted conclusion from a guess
| - A bubble map of fossil-fuel power: which statement is true
- The minimum prompt sequence to clean a table and chart it
| - Combine frequency, duration and recency to explain a recommendation
- Honest versus cherry-picked reading of the same data
|
| AI project cycleAll four AI skillsscoping and deployment not assessed | | - The self-learning robot stands in for the modelling and training stages
| - Clean and chart a messy dataset (the data stage)
- A resume-rating bias study, then retrain on balanced data (evaluation and fix)
| - Find the root cause of a failure and choose a proportional fix
|
| Directing AI / promptingPrompt EngineeringPre-Prompt Engineering at G3-4 | - Which sentence makes the chatbot draw the right picture
- Pick the clearest instruction for a birthday-card AI
| - Choose the prompt most likely to produce the wanted poster
- Reverse-engineer the steps behind a shown output
| - Spot the vague word in a climate-change prompt
- Why a chatbot stays encouraging and on-subject: its rules
| - Spot the redundant step in a prompt sequence
- Rebuild a sequence from its intermediate outputs
|
| Bias, ethics and fairnessAI Limitations & Ethics | - An AI can't fetch a toy and lacks your context: choose the safe option
| - High-stakes versus low-stakes AI errors: where a mistake really matters
| - Recreating a copyrighted photo with AI: right or wrong, and why
- Resume-rating bias by gender: which claim the chart supports
| - Face-recognition gaps across groups: find the root cause and a proportional fix
|
| Digital Literacy |
| Internet safety, passwords, authenticationUsing Digital Tools | - A pop-up says you won a bicycle you never entered for: what's safe to do
- CAPTCHA-type checks; put the steps in the right order
| - 'bl@ckSK' fails the password rules: the smallest change that passes
- Give the right people view, comment or edit access
| - 2FA: what you enter the next day (same login, fresh OTP)
- Cloud versus local storage trade-offs
| - Email to, cc and bcc visibility; respond to a breach when a password is reused
|
| Data privacy and informed consentUsing Digital ToolsInformation & Media Literacy | - Which personal details are unsafe to share with strangers online
| - Match file-sharing permissions to four different people
| - Why personalised ads appear: how browsing data is shared between services
| - Sharing with one site can pass your data to others; cookie and tracking awareness
|
| Media and information literacyInformation & Media Literacy | - Spot the clickbait headline that hides what the article is about
| - Which photo edit changes the meaning of the picture
- Spot likely AI-generated content
| - Four headlines on a phone policy: which one shows bias
- A class chat: which reactions count as bullying
- A statistically implausible claim
| - A media report cherry-picks a study: choose the honest summary
|