Games + AI

AI that drives the game. AI that helps build it. Both handled with care.

Most "AI for kids" courses are slideshows. Here, AI helps you build the games. You write the brain that can control the environment and the player: Python code is used to build the arena, makes decisions and remember what happened. Learn how to build with AI, carefully: a precise request that changes the real game, tested and run by the mentor, tested by the game, then tested by you. Knowing how far to trust AI - and how to check it - is the skill.

AI that drives the game · you write it

A brain that decides, remembers and adapts.

From lesson 1, your robot, fighter or player is driven by Python functions you write. It sees the world (positions, distances, armour, what the other players are doing), decides what to do, and remembers. That's a state machine, the same technique behind commercial games.

Write a brain that chases, flanks, escapes when armour is low, or any strategy you can think of
Beat the computer's own brains, from Easy to Expert
Write code that adapts to the environment
Decide your player's logic

# A fighter brain that reads the opponent (Fight Lab) def brain(me, enemy): if enemy is None: return 0, 0, "" d = me.distance_to(enemy) if enemy.attacking and enemy.move == "low": return 0, 0, "low_block" if enemy.attacking: return 0, 1, "" # sidestep the attack if me.special_ready and d < 1.5: return 0, 0, "special" if d > 1.3: return 1, 0, "" # close the gap return 0, 0, "punch"

AI that helps build the game · used carefully

Change the real game with AI. Then check its work.

Every member can ask for changes to their character, the environment, or the rules of the game. The request goes through the mentor, a test suite and a review before anyone uses it. That process is part of the lesson.

The safe AI loop: idea, a request with no personal info, a prediction, an adult check where the mentor runs the AI, then a test to find the AI's mistakes — The safe AI loop. Every AI change starts with an idea and ends with a test, and an adult checks it in the middle.

What the lessons teach about AI

Mapped to what schools and exam boards now expect.

Learning outcomes are tracked in the games. They appear in every lesson, and every mission is recorded on the learner card.

Safe use of AI (AISAFE)

Write an AI request with no personal information and clear limits.

England KS3: use technology safely, respectfully, responsibly and securely, protecting online identity and privacy · KS4: how changes in technology affect safety · DfE: no personal data in generative AI tools making consideration for legal, social and ethical consequences.

Checking AI code (AIREVIEW)

Check AI-written code does what was asked, test it, explain it and say where AI helped.

DfE: AI content needs critical judgement to check accuracy · JCQ: verify and acknowledge AI use (tool and date) · Curriculum and Assessment Review 2025: use AI effectively without becoming dependent.

Parents ask

Straight answers about AI in the club.

Does my child talk to an AI chatbot?

No. Learners never type into an AI tool. They write a structured request card inside the game. Only the teacher sends anything to the AI, from the teacher's own computer, and the card is checked for personal information first.

What AI is used, and what does it see?

Claude (made by Anthropic), run by Peter. It sees the text of the request card and the learner's username, which is a nickname or first name only, never a surname, school, email or location. It edits a copy of the game files; the game tests the result and Peter decides whether to keep it.

Isn't using AI to write code cheating?

Not when it's done the way exam boards say it should be: ask precisely, test the result, explain the code, and credit the tool and the date. That is exactly what the review questions and the learner card record. Lessons also make sure they can code the brain themselves before AI touches anything.

Will this help with GCSE or A-level?

The Joint Council for Qualifications now requires students to acknowledge AI use in coursework and to verify what it produces. The club practises both every week, and the computing content (variables, selection, functions, validation, testing) maps to the GCSE specifications.

Could an AI change break the game or do something unsafe?

Changes are limited to the files the teacher allows, everything is checked and anything unsafe is removed. The game's own tests also run on every change, and the teacher can undo or switch off any change. Core safety files (the brain sandbox, profiles, the AI pipeline itself) are off limits and not accessible.

The taster includes a first AI request.

Your child asks for a change to their game; you watch what happens next.

Book a free taster