Scope & sequence · grades K-12

One idea, taught for thirteen years

From Sand to Agents — K-12 Artificial Intelligence follows a single spiraled throughline. Nothing taught early is ever undone — each band revisits the same core ideas with more depth and more rigor.

silicon → chips → machine learning → neural networks → modern LLMs → prompting, caching & agents → ethics, IP, safety & careers

K-2  Smart Machines Around Us

authored

What is a smart machine? (notice & decide)

UnitTitleFocusSample VA CS SOL
U1Smart Machines Around UsMachines that 'notice and decide' vs. machines that just do their job.K.4K.52.52.13
U2Patterns EverywherePatterns let us predict what comes next — the everyday face of how machines guess.1.12.11.112.11
U3Teach the MachineMachines learn from the examples we show them; bad examples make wrong guesses.K.42.5K.92.11
U4Being Fair and Safe with Smart MachinesFair examples, private information, and honesty — people decide what's right.K.82.10K.102.142.6

3-5  How Computers Think

planned

From switches to instructions (the machine underneath, gently)

UnitTitleFocusSample VA CS SOL
U1Inside the Box: Switches and SignalsEverything a computer does is built from tiny on/off switches.mapped at authoring
U2Algorithms: Recipes a Computer FollowsStep-by-step instructions, loops, and debugging.mapped at authoring
U3Data, Sorting, and Smart GuessesComputers find patterns in data to make predictions.mapped at authoring
U4Good Data, Fair ResultsWhere data comes from, why it can be biased, and responsible use.mapped at authoring

6-8  The Machine Underneath

planned

Machine learning & neural networks (how it actually learns)

UnitTitleFocusSample VA CS SOL
U1From Silicon to ChipsTransistors, logic gates, and how hardware computes.mapped at authoring
U2What Machine Learning Really IsTraining data, models, features, and generalization.mapped at authoring
U3Neural Networks by HandNeurons, weights, and layers — an unplugged-to-code build.mapped at authoring
U4Bias, Data Quality, and EvaluationMeasuring whether a model is good — and fair.mapped at authoring
U5Security, Privacy, and Digital CitizenshipProtecting data and using systems responsibly.mapped at authoring

9-12  Modern AI: How It Actually Works

planned

Modern LLMs, agents, ethics, IP & careers

UnitTitleFocusSample VA CS SOL
U1Representation: Numbers, Vectors, EmbeddingsHow meaning becomes math.mapped at authoring
U2Training Deep NetworksGradient descent, loss, and overfitting.mapped at authoring
U3Transformers and AttentionThe architecture behind modern models.mapped at authoring
U4Large Language ModelsPretraining, tokens, and emergent behavior.mapped at authoring
U5Prompting and ContextGetting reliable results from a model.mapped at authoring
U6Caching, Cost, and LatencyEngineering AI systems that scale.mapped at authoring
U7Tools and AgentsModels that act: tool use, planning, and loops.mapped at authoring
U8Retrieval and GroundingConnecting models to real, current data.mapped at authoring
U9Ethics and Bias at ScaleHarm, fairness, and accountability.mapped at authoring
U10Intellectual Property and AuthorshipWho owns AI-assisted work; attribution and integrity.mapped at authoring
U11Safety, Society, and AI CareersAlignment, policy, and the jobs this creates.mapped at authoring

Planned bands (3-5, 6-8, 9-12) show the intended arc; unit titles are finalized when each band is authored. The K-2 band is complete and free to preview.