CurriculumParents GuideAI Literacy

From Scratch to AI: A 5-Year Coding Roadmap for Your Kid

Michael Murr··13 min read

Last updated: June 2026

The five-year arc from Scratch to AI tools is the most reliable path I have seen in 20+ years of teaching kids to code. It produces 13-year-olds who can direct AI to build real software for them. Scratch is a visual block-based programming language built by MIT for kids ages 8-16, and it removes typing friction so beginners can learn logic first. Python is a beginner-friendly text-based programming language widely used in real-world software, data, and AI work. The kid who starts with Scratch at 8 and ends with Python plus AI tools at 13 looks nothing like the kid who jumped straight to ChatGPT at 11 with no foundation. The first one is building. The second one is copying. The roadmap below is the difference.

This is the long-form view of what a five-year coding journey actually looks like, year by year. It covers what kids should be doing, what they should be building, what readiness for the next stage looks like, and where AI tools enter the picture (later than parents usually expect). It is the roadmap I would map for my own kid if I had one.

AI literacy is the ability to read, evaluate, and direct AI-generated code rather than blindly copy-paste it, and it depends on real prior coding skill.

Key Takeaways

  • The right five-year arc moves a kid from visual block coding (Scratch, ages 8-9) through Python projects (10-11) to AI literacy and AI-assisted building (12-16). Each stage builds on the previous one.
  • AI tools should enter the picture after a kid has 6-12 months of real Python under their belt. Earlier introductions produce copy-paste behaviour, not understanding.
  • The transitions between stages matter more than time spent inside any single stage. A botched Scratch-to-Python transition can lose 6 months of progress; a well-handled one accelerates the whole timeline.
  • Year 5 is where the magic happens: a 13- or 14-year-old who has followed the full arc can build complete software projects with AI tools that would have taken a professional weeks to ship five years ago.
  • This is not a rigid timeline. Some kids compress it into 3 years; some stretch it to 6. The order is what matters, not the calendar.

Table of Contents

Why a 5-Year View (and Not a 6-Month One)

Most coding programmes parents look at are sold in 6-week or 6-month chunks. That framing is convenient for billing and disastrous for actual learning. A kid does not become a real coder in 6 weeks. They become a real coder over years, and the architecture of those years matters.

The five-year view changes the question parents ask. Instead of "what will my kid have built by July?" the question becomes "where do I want my kid to be at 13?" Once you fix the destination, the path back becomes obvious. For example, AI literacy at 13 requires Python at 11. Python at 11 requires real logic at 10. Real logic at 10 requires Scratch projects at 9. Scratch projects at 9 require Scratch foundations at 8.

That backwards chain is the roadmap. Skip a step and the chain breaks. Push a step too early and the kid skips understanding to keep up, which is the same as breaking the chain just more slowly.

For the shorter Scratch-to-Python segment of this arc, our Scratch to Python roadmap goes deeper on years 1-3 specifically.

Year 1: Scratch Foundations (Ages 8-9)

The first year is about building the mental model. The kid learns that programs are sequences of instructions, that loops repeat, that conditionals choose, and that variables store. They learn this visually, in Scratch, with no syntax friction.

What the kid should be doing this year:

  • Dragging blocks to make a sprite move, change costume, and react to clicks
  • Building simple games with score tracking (Catch the Apple, basic Pong)
  • Animating short stories with scene changes
  • Starting to ask "what if I tried this?" instead of always following a tutorial

Projects that work in year 1:

  • A catch game where a basket catches falling apples
  • An animated story with two characters and three scene changes
  • A maze game where the sprite has to reach the goal
  • A short quiz with three questions and a score

Sign that year 2 is coming: the kid starts opening Scratch on their own without being asked, and the projects they save have names like "my game" instead of "tutorial 3."

If your kid has not started yet, our what age should kids start coding guide covers the developmental readiness signs that matter most at this stage.

Year 2: Scratch Projects and Real Logic (Ages 9-10)

The second year is where Scratch goes from "platform I use" to "tool I build with." The kid stops following tutorials and starts designing original projects. The concepts deepen. Nested conditionals appear. Variables get used in more places. Sprite-to-sprite communication starts working.

What the kid should be doing this year:

  • Designing their own game from a one-sentence idea
  • Debugging their own Scratch projects without an adult
  • Combining multiple sprites that interact with each other
  • Using nested if-blocks and compound conditions ("if X and Y")
  • Starting to feel limits ("I wish I could save the score so it remembers")

Projects that work in year 2:

  • A two-player Pong game with score tracking
  • A platformer with multiple levels and a final boss
  • A choose-your-own-adventure with branching paths
  • A clicker game with upgrades and unlockables

Sign that year 3 is coming: "the spark" shows up. The kid can re-explain last week's project logic in their own words, unprompted. They ask questions Scratch cannot answer. They get curious about what "the real version" of coding looks like. I cover this readiness signal in depth in my how to know your kid is ready for Python guide because parents miss it constantly.

Year 3: Python Intro (Ages 10-11)

Year 3 is the bridge year. The kid moves from visual blocks to typed text, and the structure of the move matters enormously. Done well, year 3 is the easiest transition in the whole arc. Done badly, it kills momentum for months.

The first three months of Python should not be "leave Scratch behind, start over." They should be "here is something you already built in Scratch; here is the same thing in Python." Side by side. The kid sees repeat 10 become for i in range(10): and the recognition lands within minutes. That recognition is the bridge.

What the kid should be doing this year:

  • Translating their own Scratch projects into Python line by line in early sessions
  • Writing their first standalone Python programs from a blank file
  • Learning to read error messages and parse what went wrong
  • Building text-based games and simple visual projects with Python's turtle module

Projects that work in year 3:

  • Number guessing game
  • Mad libs story generator
  • Turtle drawings (squares, stars, fractal trees)
  • Simple text adventure with three rooms

Sign that year 4 is coming: the kid stops needing the Scratch comparison. They write Python directly, from their own ideas, without having to translate from blocks in their head first. That cognitive shift usually happens in months 4-6 of Python.

For the longer Python project list at this stage, our 10 Python projects kids will actually want to build covers the specific projects that consistently land.

Year 4: Python Projects and Early AI Literacy (Ages 11-13)

Year 4 is where Python starts feeling powerful. The kid is no longer translating from Scratch. They are writing Python natively, with their own ideas, building things that take multiple sessions and earn real "I built this" pride. This is also the year AI tools enter the picture for the first time, carefully.

What the kid should be doing this year:

  • Building multi-file Python projects with functions and basic OOP
  • Writing programs that read and write files, or talk to APIs
  • Using pygame for graphical games with sprites, music, and scoring
  • Trying their first AI-assisted project, with a tutor watching how they use the tool
  • Starting to evaluate AI outputs critically instead of accepting them blindly

Projects that work in year 4:

  • A pygame game with sprites, sound, and a scoring system
  • A web scraper that pulls live sports scores or game prices
  • A chatbot powered by a free LLM API with a custom personality
  • A small CLI tool that solves a real problem the kid has (a homework reminder, a chore tracker)

Where AI fits in year 4: AI tools are introduced, not relied on. The kid uses AI to explain Python concepts, debug code, and occasionally generate small snippets they then read and modify. They are not using AI to write entire projects yet. The reading comprehension has to come first.

I wrote about what happens when AI tools are introduced too early without this Python foundation in can kids learn AI: what parents need to know.

Sign that year 5 is coming: the kid can read AI-generated Python code, identify what is wrong with it, and fix it. That comprehension threshold is the gateway to year 5.

Year 5: Advanced Python and AI Tools (Ages 13-16)

Year 5 is the payoff year. The kid has 4 years of solid foundation, real Python literacy, and the maturity to use AI tools as a collaborator instead of a crutch. This is where the arc starts producing results that look genuinely impressive to adults.

What the kid should be doing this year:

  • Building complete, multi-week software projects (games, tools, simple web apps)
  • Directing AI tools like Claude Code or Cursor to write Python for them, then reading, testing, and fixing what comes back
  • Decomposing real problems into smaller engineering tasks
  • Specialising into a track they care about (games, web, data, ML)

Projects that work in year 5:

  • A complete pygame project shipped end to end (menu, multiple levels, save state, music)
  • A simple web app with Flask or FastAPI that does something real
  • An AI-assisted personal project decided and led entirely by the kid
  • A first machine learning project using scikit-learn on a real dataset

The Stacy story. Earlier this year I gave an 11-year-old student of mine a $20 Claude Code subscription. She is named Stacy. She had been struggling with AI tools for a while. It was the kind of struggle where she could use ChatGPT but never quite built anything from start to finish with it. We turned it on. Within the month she had built a complete game by herself, end to end, directing Claude Code through the design and implementation. That moment is what year 5 looks like, just compressed. A kid with real Python literacy unlocks the AI tool and goes from "I want to make this" to "I made this." The full story is in my I gave an 11-year-old Claude Code post.

By the end of year 5, the kid has options no previous generation of teenager had. They can ship software. They can use AI to multiply their output. They can think clearly about how to build things. That is the destination the five-year arc earns.

The 5-Year Roadmap at a Glance

YearAgeSkill LevelMain ToolsExample ProjectsSign of Readiness for Next
18-9Block-coding beginnerScratchCatch game, animated storyOpens Scratch unprompted, builds own ideas
29-10Scratch builderScratchPong, platformer, choose-your-adventure"The spark", asks Scratch-limit questions
310-11Early PythonPython (turtle, basic)Number guess, Mad Libs, turtle artWrites Python from blank file, no Scratch comparison
411-13Python projects + AI introPython, pygame, requests, LLM APIsPygame games, chatbots, scrapersReads AI code, debugs it, fixes it
513-16Advanced Python + AI directionPython, frameworks, Claude Code/CursorComplete apps, AI-led projects, MLShips real software they imagined

Where Most Parents Get the Timing Wrong

Five mistakes I see consistently, all of them recoverable if caught early.

Mistake 1: Skipping Scratch because "it's for little kids." Scratch is not babyish. It is a logic-foundation tool that produces faster Python learners 12 months later. Skipping it usually costs more time than it saves. A kid who skipped Scratch typically spends month 1-3 of Python fighting syntax instead of learning concepts, then plateaus.

Mistake 2: Pushing Python before the spark. Parents see another kid moving to Python and decide their child should too. Without the readiness signs, the kid spends 2-3 months frustrated with syntax errors, then quietly disengages. Recovery from that frustration takes longer than the patience would have.

Mistake 3: Introducing AI too early. A 9-year-old with ChatGPT does not learn coding faster. They learn to copy-paste faster. AI tools belong in year 4 at the earliest, after enough Python literacy to read what the AI produces.

Mistake 4: Believing the 6-month app promise. Some platforms advertise that your 9-year-old will build a real app in two months. They will not. A 9-year-old at month 2 is still on if-statements. Real applications come in years 4-5. Schools that promise otherwise are inflating expectations they cannot deliver on. As a result, parents who buy in end up burned. I cover this in detail in my piece on coding schools promising 9-year-olds will build apps.

Mistake 5: Inconsistent practice. Two months on, three months off, six weeks on, six weeks off. Coding builds on itself. Gaps cost more than the bursts add. Year 1 of this roadmap requires roughly one focused hour per week, every week. That consistency is the architecture.

Frequently Asked Questions

Does my kid really need 5 years to learn to code? Not necessarily. Some kids compress this arc into 3 years with high intensity. Others stretch it across 6. The order matters, not the calendar. Five years is the right framing because it sets the destination far enough out that the architecture has time to work. Faster is fine if the foundation is real.

At what age should my kid start using AI tools like ChatGPT or Claude? For coding specifically, year 4 is the earliest, around ages 11-13, after 6-12 months of real Python. For general use (asking questions, writing help with school) the conversation is different and depends on the family. However, for coding, premature AI use produces copy-paste behaviour that is genuinely harder to undo than to prevent.

What if my kid is already 12 and has never coded? The arc compresses. A 12-year-old with no coding background can typically do 6 months of Scratch, 12 months of Python, and arrive in the AI-ready zone by 14. Older beginners actually move faster than younger ones in the early stages. Their abstract thinking is more developed. The destination is the same; the path is shorter.

Can my kid skip Scratch and start with Python? Sometimes, usually for kids 13 and up or for kids with strong logic foundations from chess, math, or another structured discipline. However, for most 8-to-11-year-olds, skipping Scratch trades 1 month of perceived saved time for 6 months of fighting Python syntax without a mental model. The arithmetic does not favour skipping.

What about JavaScript, C++, or other languages? Python is the strongest first text-based language for kids because of its readable syntax and the size of its educational ecosystem. Meanwhile, JavaScript becomes useful around year 5 if the kid wants web development. C++ is rarely the right choice before 15-16 because the syntax overhead distracts from concept learning. The roadmap above stays in Python intentionally. Other languages can layer in once Python is solid.

How much does the 5-year arc cost? At the low end, almost nothing, Scratch is free, Python is free, and self-directed kids can go far on YouTube. At the high end, with 1-on-1 tutoring throughout, the investment is real. The decision usually comes down to whether the kid is self-directed enough to make the free path work. Most kids in the 8-11 range are not. Therefore, the middle path is structured tutoring at key transitions (Scratch-to-Python, Python-to-AI) and lighter support in between.

The Bottom Line

The kid who follows the five-year arc from Scratch to AI does not just "know how to code" by 13. They know how to think about software, how to use AI as a collaborator, and how to build things that matter to them. That destination is worth designing for, not just stumbling toward.

Want a tutor's read on where your kid is on the 5-year arc? Book a free Discovery Call and we will map their current stage, what the next step looks like, and how to get there without the timing mistakes that cost most kids 6-12 months.

Enjoyed this article?

Your child can learn this and more with a dedicated 1-on-1 tutor.

Book a Free Discovery Call