The old model
- AI treated as a topic inside computing.
- Students learn what AI is, but rarely use it to build serious work.
- Non-technical students feel excluded before they begin.
- Curriculum cycles move slower than the tools themselves.
Cohort 2026 · enrolment open · North West UK pilot
Full-Spectrum AI Innovation Academy · Ages 6–19+
AI across every subject.
Real projects in every hand.
ARTIZAI is an after-school academy where students aged 6 to 19+ use professional AI tools — AlphaFold, Midjourney, Fusion 360, Whisper, Claude, Roboflow — to go deeper into science, maths, arts, languages, humanities, social science, and physical making.
The idea
AI is not a subject. It is a new cognitive instrument.
Students do not come to ARTIZAI to "learn AI" in isolation. They use AI as a research partner, design partner, language coach, studio collaborator, simulation engine, and making assistant inside the disciplines they already care about.
Every sprint ends with evidence: a paper, a prototype, a bilingual edition, a policy brief, a film, a drone, a working sensor, or another portfolio-ready output.
Different by design
From one computing lesson to a whole academy of applied AI.
The ARTIZAI model
- AI embedded across every subject and age tier.
- Students produce visible outputs in 12-week project sprints.
- Craft, judgement, ethics, and domain knowledge stay central.
- Tool modules update as AI models and professional workflows evolve.
Seven ARTIZAI Labs
A complete after-school offer, not a single-track STEM club.
Each lab uses the same sprint architecture, but changes the domain, tools, mentors, and output. Students enter through the subject they love and leave with transferable AI fluency.
Science Lab
Biology, chemistry, physics, environmental data, protein structures, sensors, and AI-assisted scientific papers.
- AlphaFold-style protein modelling
- AI microscopy and field data analysis
- Environmental sensor builds
Mathematics Lab
Pure and applied maths where intuition meets symbolic computation, proof assistants, simulations, and data science.
- Fractals and visualisation
- Traffic and city data models
- AI tutoring prototypes
Arts Lab
Visual art, music, drama, design, film, and AI as a serious creative collaborator with human craft finishing.
- Short films from script to score
- AI-assisted identity systems
- Music composed and performed live
Social Science Lab
Economics, psychology, sociology, geography, policy briefs, public datasets, and responsible data storytelling.
- Levelling-up data analysis
- AI bias and perception studies
- Regional sustainability models
Languages Lab
Modern languages, linguistics, translation, voice tools, multilingual chatbots, and global communication projects.
- Bilingual editions with commentary
- Museum and heritage chatbots
- Podcast transcription and production
Humanities Lab
History, philosophy, ethics, literature, primary source analysis, digital humanities, and AI debate partners.
- Historical source pattern analysis
- AI ethics essays and critiques
- Literary network visualisations
Making & Engineering Lab
Drones, vehicles, e-mobility, electronics, CAD, British craft, and the physical proof that AI literacy is a making skill.
- FPV drone airframes and flight systems
- 1:10 electric vehicles and coachbuilding
- Custom PCBs, motors, and telemetry
Subject coverage · 14 disciplines · every brief click-to-open
How AI helps each subject become a lab.
From Chemistry to Ethics, every discipline has a full lab brief: classical method vs AI-augmented method, the actual AI tools we run (with direct links), three concrete classroom projects, real physical instruments, and career pathways. Tap any card.
⚛ STEM · 5
◐ Creative · 2
∮ Social · 4
⌛ Humanities · 3
Every brief is updated quarterly as new AI platforms ship. Match a subject to your child →
Seven virtual AI studios · classical lab + AI co-scientist
Real instruments, real datasets, real AI tools — running side by side.
Every lab keeps the microscope, soldering iron, sketchbook, or proof-pad — and pairs it with the AI tools actual scientists, engineers, artists and policy researchers use today. Below: tool stack, classroom-tested scenarios, and direct links to every platform we run.
// the_artizai.method
The ARTIZAI Scientific Method
Four-phase loop · 12-week sprint · every lab · every age tier
-
PHASE 01 · WEEKS 1–2 · hypothesis
Observation & Question
Students begin with a real question in the subject — observable, falsifiable, anchored in a primary source or live measurement, not a worksheet.
output → research brief + hypothesis -
PHASE 02 · WEEKS 3–5 · investigation
AI-Augmented Investigation
Students use AI to search literature, compare models, run simulations, translate sources, generate counterexamples, and challenge their own assumptions.
output → annotated dataset + prompt trail -
PHASE 03 · WEEKS 6–8 · peer review
Critique & Iteration
Small live cohorts critique the work, defend assumptions against AI-generated counter-arguments, surface bias, and plan the next iteration.
output → revised method + iteration plan -
PHASE 04 · WEEKS 9–12 · publication
Build & Publish
Each sprint ends with a portfolio artefact: a paper, prototype, drone, dataset, film, or live performance — versioned, peer-reviewed, and shipped.
output → portfolio entry · Bronze · Silver · Gold -
LOOP · NEXT SPRINT · iterate
Findings, failures, and questions from publication feed Phase 01 of the next sprint — the academy compounds knowledge, not minutes.
Science Lab · Chemistry · Biology · Physics · Climate
From wet-bench experiments to AI-augmented research
Students keep the microscope, conical flask, pipette, and oscilloscope — but pair every measurement with an AI research partner that explains, predicts, and challenges results.
Classical lab instruments
- Compound + USB digital microscope, slide kits, stains
- pH meter, conductivity probe, spectrophotometer (Vernier / PASCO)
- Bunsen, conical flask, burette, balance, hotplate-stirrer
- Oscilloscope, function generator, ray-box, force-meter
- Arduino / micro:bit + DHT22, PMS5003, BME680 sensors
AI co-scientist (new)
- Protein structure prediction from sequence (AlphaFold)
- Reaction pathway + spectrum interpretation (GPT-4 / Claude)
- Microscopy image classification (Roboflow + custom CNN)
- Live PhET simulations cross-checked with sensor data
- Species ID from photos + audio (iNaturalist, BirdNET)
- AlphaFold DB
- PhET Sims
- Labster
- Foldit
- iNaturalist
- BirdNET
- Semantic Scholar
- Wolfram Alpha
- Galaxy (bioinformatics)
- Roboflow Vision
Real classroom scenarios
- Air quality across our school site — students deploy 5 PMS5003 + BME680 nodes on ESP32 boards, log PM2.5 every minute to a Google Sheet, then ask Claude to fit a diffusion model and flag the worst micro-zone before assembly.
- Predict a protein, then 3D-print it — feed a novel sequence to AlphaFold via the EBI API, export the .pdb, render in PyMOL, print on a Prusa MK4 at 0.12 mm, label the binding pocket on a scientific poster.
- Bioplastic from kitchen starch — synthesise four variants (cornflour, agar, gelatin, alginate), test tensile strength on a hand rig, ask GPT to compare reaction yields, publish results as a Year-10 micro-paper.
Mathematics Lab · Pure · Applied · Data Science · Proof
Maths becomes a working laboratory, not a worksheet factory
Students use AI to test intuition, generate counterexamples, run real simulations, and explain proofs — all anchored to Python notebooks, real datasets, and proof assistants.
Classical maths kit
- Graphing calculator, compass, protractor, set square
- Pen-and-paper proof, geometry construction, mental arithmetic drills
- Statistical tables, log books, lab notebook for hypotheses
AI maths bench (new)
- Symbolic computation + step-by-step (Wolfram Alpha, SymPy)
- Interactive dynamic geometry (GeoGebra, Desmos)
- Proof assistant with AI hints (Lean + GitHub Copilot)
- Cloud notebooks for data + ML (Google Colab, Kaggle)
- AI tutor that demands reasoning first (Khanmigo)
- Wolfram Alpha
- GeoGebra
- Desmos
- Khanmigo
- Google Colab
- Kaggle Learn
- Lean Prover
- Mathpix OCR
- Manim
- GitHub Copilot
Real classroom scenarios
- Traffic around the school gates — students count cars at 3 entrances for one week, load CSV into Colab, fit a Poisson model with SymPy, build an agent-based simulation in Mesa, and present a redesigned drop-off zone to the head teacher.
- Premier League prediction sprint — pull match data from the Football-Data API on Kaggle, train a logistic regression with scikit-learn, ask Claude to critique feature choices, and submit weekly forecasts against a leaderboard.
- Prove it in Lean — Year-12 students formalise the proof that √2 is irrational in Lean 4, use Copilot for tactic hints, and write a Manim animation explaining each step to younger students.
- Fractal museum — generate the Mandelbrot and Julia sets in a Colab notebook, render with Manim, and 3D-print a Menger sponge on the MK4 — younger students operate the zoom controls in GeoGebra.
Arts Lab · Visual · Music · Film · Drama · Design
AI as creative collaborator — human taste, craft and authorship stay central
Students learn to brief, curate, remix, perform, and finish work with human judgement. Every output ships with a documented prompt trail and copyright audit.
Classical studio kit
- Sketchbooks, oils, acrylics, screen-print, lino, etching plates
- DSLR + tripod, boom mic, lavalier, audio interface
- Acoustic + electric instruments, MIDI keyboard, drum pads
- Wacom tablet, lightbox, darkroom, model-making tools
AI creative bench (new)
- Image generation + editing (Midjourney, Adobe Firefly, Stable Diffusion)
- Music composition + stems (Suno, Udio, Magenta)
- Video + motion (Runway Gen-3, Pika)
- Voice + dubbing (ElevenLabs, Descript)
- Script critique + dramaturgy (Claude, GPT-4)
Real classroom scenarios
- "My Town in 60 seconds" film — students write a script with Claude, generate storyboard frames in Midjourney, shoot live B-roll on a DSLR, score the music with Suno, voice the narration in ElevenLabs, and edit in DaVinci Resolve. Every prompt and edit is logged.
- Public AI exhibition — pair each artwork with its full prompt trail, seed values, human paint-over photographs, and a 100-word authorship statement — hung side-by-side in the school gallery.
- Live concert with AI — compose three motifs in Magenta Studio, perform live on guitar and cello while real-time stems play, document the human-AI handoff.
Social Science Lab · Economics · Psychology · Geography · Policy
Real datasets, real surveys, real civic stakeholders
Students learn to read public data, design rigorous surveys, audit AI for bias, and translate findings into journalism, policy briefs, and interactive dashboards.
Classical methods
- Pen-and-paper survey design, observational fieldwork
- Census and ONS data on printed gazettes
- Excel pivot tables and hand-drawn cartography
- Stakeholder interviews, focus groups, ethnographic notes
AI civic bench (new)
- Interactive UK + global data explorers (ONS, OWID, Earth Engine)
- Survey design + bias audit (Claude / GPT-4 + Qualtrics)
- One-click chart and map builders (Datawrapper, Flourish)
- Scenario simulation in agent-based models (NetLogo, Mesa)
- Live dashboards (Streamlit, Observable)
- UK ONS
- Our World in Data
- Flourish
- Datawrapper
- Google Earth Engine
- Streamlit
- Observable
- NetLogo
- Claude (bias audit)
Real classroom scenarios
- Levelling-up dashboard — pull IMD 2024 + ONS census data, ask Claude to suggest weighting schemes, audit them for bias, build a Streamlit dashboard, present a memo to the local councillor.
- AI media trust experiment — show 60 peers four versions of the same news story (human, AI-edited, AI-generated, AI-translated), gather Likert responses in Qualtrics, fit a mixed-effects model in R.
- Climate-risk map of our city — overlay Environment Agency flood data on Earth Engine, identify three vulnerable streets, propose mitigation in a policy brief.
Languages Lab · Modern Languages · Linguistics · Translation
AI as conversation partner, dialect coach, and corpus analyst
Language is culture, power and memory. Students compare AI translations, study real linguistic corpora, and produce bilingual editions, podcasts and museum guides.
Classical methods
- Printed grammar primers, dictionaries, phrase books
- Listening from cassette / CD, recitation, role-play
- Pen-pal letters, immersion week, language exchange
- Linguistic fieldwork notebooks, IPA transcription
AI language bench (new)
- Conversation simulator + pronunciation coach (Speak, Duolingo Max)
- Side-by-side translation critique (DeepL vs GPT vs Claude)
- Speech-to-text + dialect transcription (OpenAI Whisper)
- Multilingual avatar dubbing (HeyGen)
- Corpus linguistics + collocation analysis (Sketch Engine)
Real classroom scenarios
- Museum multilingual guide — Year-11 students interview a curator, transcribe in Whisper, translate to EN/DE/UR/CY/PL, voice with ElevenLabs, deploy a QR-triggered audio guide for a real local museum.
- Three translators, one poem — feed the same Heaney poem to DeepL, GPT-4 and Claude, annotate where each loses idiom, register, or rhythm, and publish an annotated bilingual edition.
- Pronunciation lab — record a French dialogue, run it through Whisper, compare F0 pitch contours in Praat against a native speaker, drill the worst phonemes with Speak.
Humanities Lab · History · Philosophy · Ethics · Literature
Read more sources, argue more carefully, see patterns AI can't
The human task remains judgement: what matters, what is missing, what is true. AI helps with the volume of reading; humans keep authority over the argument.
Classical methods
- Archive visits, microfilm, palaeography, marginalia
- Close reading, essay drafting, Socratic seminar
- Footnoted bibliography, primary vs secondary sourcing
- Oral history interviews on cassette / minidisc
AI humanities bench (new)
- Handwritten text recognition for archives (Transkribus)
- NotebookLM podcast briefs from primary sources
- Text mining + topic models (Voyant, BookNLP)
- Character + citation networks (Gephi)
- Argument critique + devil's-advocate (Claude)
Real classroom scenarios
- Suffrage archive exhibit — students transcribe 487 handwritten letters with Transkribus, cluster themes in Voyant, ask Claude to summarise, then publicly debunk three AI summaries that disagree with the primary sources.
- Middlemarch character network — feed the Gutenberg text through BookNLP, build a Gephi network of character interactions, present a 10-minute reading of the most-connected scene.
- AI as devil's advocate — Year-12 philosophy essay drafted with Claude attacking every claim, footnoted with primary citations the AI cannot fabricate.
Making & Engineering · Drones · Vehicles · Robotics · CAD · British Craft
AI designs and diagnoses — humans build, fly, test and refine
ARTIZAI's physical signature. Generative design, simulation, computer vision and AI tuning meet real motors, real PCBs, real props, and the British workshop tradition.
Classical workshop kit
- Soldering station, multimeter, oscilloscope, bench PSU
- Lathe, mill, drill press, bandsaw, hand tools, jigs
- 3D printer (Prusa MK4 / Bambu), laser cutter, CNC router
- Drone parts: F7 flight controller, ESCs, BLDC motors, 4S LiPo
- Pixhawk, Crazyflie, Raspberry Pi, ESP32, IMU, GPS, FPV camera
AI engineering bench (new)
- Generative design + topology optimisation (Fusion 360, nTop)
- CFD + FEA simulation (SimScale, Ansys Discovery)
- PCB layout + DRC review (KiCad + AI assistant)
- Flight controller tuning (Betaflight, ArduPilot, PX4)
- Computer vision for aerial mapping (Roboflow, WebODM, OpenCV)
- Robotics sim (NVIDIA Isaac Sim, Gazebo)
- Fusion 360
- Onshape
- FreeCAD
- Tinkercad
- KiCad
- SimScale
- Betaflight
- ArduPilot
- PX4
- Roboflow
- OpenDroneMap
- Gazebo Sim
- NVIDIA Isaac
Real drone & engineering scenarios
- Lightweight FPV drone arm — model the arm in Fusion 360, run topology optimisation with a 4N load case + 0.3 safety factor, print the AI-optimised version in PLA-CF, stress-test both designs to failure on a hand jig, log the mass saving (target: −18%).
- Crop / playground mapping mission — fly a pre-planned grid mission in Mission Planner with a Pixhawk-based quad, post-process the 280 photos in WebODM to produce an orthomosaic + DSM, train a Roboflow model to detect litter / weeds, deliver a heatmap to the school groundskeeper.
- Search-and-rescue object detection — collect aerial footage of dummy "missing person" targets in school grounds (with permission), label 600 frames in Roboflow, train YOLOv8, run inference on the Raspberry Pi 5 companion computer in flight.
- PID tuning the scientific way — log Betaflight blackbox flights, plot step responses in Python, ask Claude to suggest P / I / D adjustments and explain the reasoning, A/B test in calm conditions, document gains.
- 1:10 EV chassis — generative-design a chassis plate in Fusion, CFD-test airflow in SimScale, hand-form the aluminium body panel in the workshop, log telemetry from a custom ESP32 board over MQTT.
- Autonomous indoor flight — train a small reinforcement-learning policy in NVIDIA Isaac Sim, transfer to a Crazyflie 2.1 in the school hall, race against a hand-piloted version.
12-week sprint
A repeatable structure that still leaves room for invention.
Define & Research
Students use AI as a research partner to shape the brief, interrogate assumptions, and find the right domain questions.
AI-Assisted Design
Options are generated, curated, criticised, and checked by mentors for feasibility, originality, and safety.
Prototype & Build
The main investigation or fabrication phase: code, craft, electronics, writing, performance, analysis, or fieldwork.
Iterate & Showcase
Results are tested, improved with AI feedback, presented publicly, and documented in the student's portfolio.
Ages 6 to 19+
One continuous learning chain from curiosity to professional-grade work.
The portfolio grows with the student, giving universities, employers, apprenticeship recruiters, and parents a clear record of progress across years.
Personalised learning plan · in under 3 minutes
Tell us about your child. We build a plan, an assessment, and a daily timetable.
Six short steps. Inspired by mastery-based programmes like Alpha School, but built around the seven ARTIZAI Labs, real AI tools, and physical making. Output: a 2-hour daily plan, a weekend project, a 6-week summer track, and a recommended starter tool kit — saved locally and exportable.
Teacher CPD · in-school or virtual
ARTIZAI Faculty Programme — train the teachers who will run the labs.
A specialist CPD pathway for classroom teachers, heads of department, and SLT. Choose in-school residency (we send a faculty pair to your site for 5 days) or virtual cohort (live evening seminars + an online studio). Both end in an accredited ARTIZAI Practitioner badge and a ready-to-run sprint kit.
School residency
- 2 ARTIZAI faculty on-site for 5 consecutive days
- Live demo lessons with your real students
- Hardware kit installed: drones, sensors, 3D printer, microscope cam
- Walkthrough of safeguarding, AUP, and the AI Acceptable-Use policy
- Twilight CPD session for the whole staff body
- 30-day post-residency Slack/Teams support
5 days on-site · up to 20 staff trained · cohort badge issued at end
Cloud Studio cohort
- 8 live evening seminars (Tue + Thu · 18:30–20:00 UK time)
- Cloud Studio access for AlphaFold, Roboflow, KiCad, Colab projects
- Optional Saturday studio (10:00–13:00) for build-along sessions
- Recorded back-catalogue + transcribed seminars
- Take-home loan kit posted UK-wide (sensors, micro:bit, micro-drone)
- Final showcase: every teacher runs one real sprint in their own class
8 weeks · 16 live hours + studio time · individual or department signup
Eight modules · each one ends with a classroom-ready artefact
AI literacy for teachers
Tokens, training data, hallucination, alignment, costs. The end-of-week test the press keeps failing.
Output: 1-page school-wide AI explainer.
Prompt craft & AI tutoring
Use Claude / ChatGPT / Khanmigo as a teaching assistant — not a cheat sheet.
Output: 5 reusable lesson prompts per subject.
Safeguarding & ethics
UK Online Safety Act, ICO guidance, DfE EdTech, bias, data minimisation, consent.
Output: department AUP + AI risk assessment.
Subject-specific AI tools
Branch by subject — Science (AlphaFold, PhET), Maths (Wolfram, Lean), Arts (Firefly, Suno), etc.
Output: 3 AI-augmented lesson plans in your subject.
Sprint design
Run a 12-week sprint from a real question to a public outcome with portfolio evidence.
Output: full 12-week sprint scheme of work.
Making & physical kit
Arduino, micro:bit, 3D printing, soldering, drones — safety, procurement, lesson flow.
Output: kit inventory + 6 maker mini-projects.
Assessment & portfolio
Bronze / Silver / Gold criteria, rubrics, evidence capture, mentor critique pods.
Output: tier-mapped assessment rubric.
Live classroom sprint
Run one real sprint with your class while ARTIZAI faculty observe and coach.
Output: filmed sprint + reflective practitioner essay.
Who this is for
- Subject teachers (any of the 7 ARTIZAI labs)
- Heads of department designing AI strategy
- Computing & D&T leads expanding their offer
- SLT & MAT leaders making procurement decisions
- FE college and sixth-form lecturers
- Trainee teachers on PGCE / Teach First / School Direct
What you take away
- ARTIZAI Practitioner badge — accredited CPD certificate
- Full sprint pack — schemes of work, prompt library, rubrics
- Hardware starter kit — micro:bit + sensors + micro-drone (loan or buy)
- Cloud Studio access — 12 months of AI tool credits for your class
- Network — Slack/Teams channel with the UK practitioner cohort
- Renewal path — Advanced and Lead Practitioner badges
Funding routes
This is an eligible CPD spend.
- DfE Computing Hubs
- National Centre for Computing Education
- Teacher Development Premium
- MAT central CPD budget
- Local authority modernisation funds
- Innovate UK BridgeAI
- Industry sponsorship (per-cohort)
Train your teachers before September.
Residencies booking now for the summer term. Virtual cohorts run rolling every 8 weeks.
The gravity well
Where AI leaves the screen and becomes a built object.
The Making & Engineering Lab gives ARTIZAI its distinctive UK signature: drones, vehicles, e-mobility, electronics, CAD, and British craft heritage taught as one integrated practice.
Students move from AI-assisted design to physical fabrication, then test, iterate, document, and present work that can be understood by industry mentors.
Drone
PCB
CAD
Craft
E-Mobility
Drone Engineering
Aerodynamics, flight controllers, topology-optimised parts, PCB design, and CAA-aware operating literacy.
Scale Vehicles
AI chassis optimisation, hand-formed body panels, telemetry, custom drivetrains, and British coachbuilding logic.
Future Mobility
Employer-linked briefs for last-mile delivery, sustainable transport, school mobility, and regional infrastructure.
British Craft
Resident artisans bring materials intelligence from textiles, ceramics, metalwork, slate, leather, and model-making.
Electronics & CAD
KiCad, Arduino, Raspberry Pi, ESP32, FreeCAD, Fusion-style workflows, soldering, sensors, and embedded systems.
Drone Engineering Bench · the UK signature lab
Students don't just fly drones. They design the airframe, lay the PCB, write the tuning code, and read the telemetry.
The ARTIZAI Drone Bench is where Maths, Physics, Computing, Engineering, and AI converge on one buildable object. Below: the full anatomy, the software stack pilots actually use, the hardware tiers, the legal pathway, and how AI changes every step from CAD to flight.
The 10 subsystems students master
- Frame — carbon plate stack, X / H / dead-cat geometry, vibration paths.
- Motors — BLDC sizing (2207, 2306), KV vs prop pitch, thrust-to-weight target ≥ 4:1.
- ESCs — DShot, BLHeli32 / AM32, current draw, thermal headroom.
- Flight controller — F7/H7, IMU, gyro filtering, blackbox logging.
- Receiver — ELRS, CRSF, link budget, fail-safe behaviour.
- VTX + camera — analog vs digital (DJI O3, Walksnail), MIPI, latency budget.
- Battery — LiPo chemistry, C-rating, sag, safe charging & storage.
- GPS & compass — for ArduPilot / PX4 autonomous modes.
- Companion compute — Raspberry Pi 5 / Jetson Orin Nano for on-board CV.
- Telemetry & mission — MAVLink over MQTT, Mission Planner, ground station.
The real software stack (every tool below is one a working pilot or engineer uses)
Hardware tiers — students progress through three buildable platforms
Tinywhoop indoor
Whoop class (65 mm), brushed motors, 1S battery. Safe to fly in the school hall.
- BetaFPV Cetus Pro / Meteor65
- F4 1S AIO board · ELRS rx
- Goggles + ExpressLRS radio (Radiomaster Pocket)
- Skills: throttle discipline, line-of-sight, fail-safe
3″–5″ freestyle FPV
Carbon X-frame, brushless 2207 motors, 4S/6S LiPo. The mainline build platform.
- iFlight Nazgul / Apex frame
- F7 stack · 45A 4-in-1 ESC · DJI O3 or analog VTX
- Soldering · Betaflight tuning · blackbox analysis
- Skills: PID tuning, generative arm design, CFD test
Autonomous quad
Pixhawk-class GPS quad with Raspberry Pi 5 / Jetson companion for AI.
- Holybro X500 / Pixhawk 6C
- ArduPilot or PX4 · GPS · LiDAR / depth cam
- MAVLink · OpenCV · YOLOv8 · OpenDroneMap
- Skills: mission planning, SLAM, computer vision, photogrammetry
Where AI changes every step of the build
Design
Fusion 360 generative design + topology optimisation cuts arm mass by 15–25% under the same 4N load case.
Simulate
SimScale / AirSim runs CFD on the airframe and rehearses flight policies in software-in-the-loop before any hardware risk.
Lay PCB
Flux AI suggests trace widths for current draw and runs DRC; KiCad exports Gerbers to OSH Park.
Tune
Claude reads Betaflight blackbox logs, plots step responses in Python, recommends P / I / D + filter changes with reasoning.
Perceive
YOLOv8 trained in Roboflow on aerial footage spots litter, people, weeds, or crop stress at 30 fps on a Pi 5.
Navigate
Reinforcement-learning policies trained in Isaac Sim transfer to a Crazyflie 2.1 — sim-to-real handoff documented.
Map
OpenDroneMap stitches 280 photos into an orthomosaic + DSM; AI flags change vs last week's flight.
Diagnose
If something breaks, GPT reads the flight log + crash photo and proposes a likely root cause (frame, motor, RX, gyro).
Eight real classroom missions — by difficulty
Tinywhoop obstacle course
Set up a course of paper gates in the school hall. Students fly increasingly hard lines, log lap times, and tune throttle expo.
Output: lap-time leaderboard + flight video.
Fail-safe drill
Deliberately walk out of receiver range to test ELRS fail-safe, document behaviour, then change settings and re-test.
Output: safety report with screenshots.
PID tuning with blackbox
Fly the same throttle profile before/after a P-gain change, export blackbox in Python, plot step response, compare overshoot.
Output: Jupyter notebook with annotated plots.
Topology-optimised arm
Model an arm in Fusion 360 with a 4N load case + 0.3 safety factor, run generative design, print both versions, fail-test on a hand jig.
Output: mass savings vs strength report (target −18%).
Custom PCB sensor pod
Design a small ESP32 + IMU + barometer board in KiCad, manufacture at OSH Park, mount under the quad, log telemetry over MQTT.
Output: PCB + live altitude dashboard.
Aerial mapping mission
Fly an auto grid with a Pixhawk + Mission Planner, capture 280+ overlapping photos, post-process in OpenDroneMap to get an orthomosaic.
Output: 5 cm/px map + DSM of school grounds.
YOLO litter detection
Label 600 aerial frames in Roboflow, train YOLOv8n, deploy on a Raspberry Pi 5 companion, run live detection in flight.
Output: heatmap + detection video for groundskeeper.
Sim-to-real RL nav
Train a small RL policy in Isaac Sim, transfer to a Crazyflie 2.1, race against a hand-piloted version in the school hall.
Output: sim transcript + real flight log + race time.
UK CAA pathway · safety first
Every pilot earns the right paperwork before flying outdoors.
The drone curriculum is mapped to UK Civil Aviation Authority categories. Students under 13 fly indoors only; from 13+ the school registers as the operator, students earn the Flyer ID, and the academy carries Operator ID, third-party insurance, and a documented risk assessment per flight.
- CAA Flyer ID
- A1 / A3 sub-categories
- Operator ID (school)
- Pre-flight risk assessment
- LiPo storage SOP
- Third-party insurance
- No-fly-zone check (Drone Assist)
Cross-lab advantage
The most interesting work happens between disciplines.
Molecular orbitals become gallery prints through Python computation and generative visual systems.
Topology-optimised drone airframes are printed, tested, and compared with conventional designs.
Students analyse how different regional-language newspapers frame the same civic issue.
Historical craft research turns into modern material experiments with AI-assisted colour and pattern matching.
Delivery model
Three ways to bring ARTIZAI into a region.
Regions can start with a flagship centre, extend through school-hosted labs, and scale through city hubs that activate underused school infrastructure after hours.
Standalone Centre
A regional anchor with ARTIZAI-owned premises, complete equipment, full facilitation, and maximum brand control. Best for cities with multiple feeder schools.
- Funding fit
- Innovate UK TechLocal, BridgeAI, corporate sponsorship
- Best use
- Pilot launch, flagship making studio, regional showcase
School-Hosted Lab
A school provides a DT room, IT room, or repurposed classroom while ARTIZAI provides curriculum, training, platform, equipment spec, and mentor support.
- Funding fit
- DfE EdTech, school capital, KTP routes
- Best use
- Lower-cost expansion into suburban, rural, or MAT settings
City Hub
One equipped host school opens its ARTIZAI lab to partner schools, Saturday intensives, holiday programmes, inter-school project days, and teacher CPD.
- Funding fit
- UKSPF, levelling-up funds, local authority modernisation
- Best use
- Scaling access without every school building a full lab
Always current
Stable pedagogy. Updating AI tools. Serious safety.
Lab frameworks
Annual review keeps the sprint model, assessment, portfolio criteria, and subject progression coherent.
AI tool modules
Six-monthly updates swap in better tools only when they improve student outcomes, safety, access, and cost.
Ethics thread
Quarterly updates track UK Online Safety Act, ICO, safeguarding, and AI Safety Institute guidance.
Alumni loop
Pioneers and certified facilitators bring new knowledge from universities, employers, and local industry.
Global proof
Similar models exist. ARTIZAI combines their strongest lessons into one UK-ready system.
The world already has serious experiments in virtual, mastery-based, project-based, and AI-personalised learning. ARTIZAI should not copy one model. It should assemble the best proven patterns around its own distinctive edge: AI across every subject plus physical making and British craft.
Sora Schools
Remote grades 6-12, interdisciplinary projects, personal mentors, college preparation, and a school day that feels closer to a studio or lab than a bell schedule.
Official siteSchool of Humanity
Online middle and high school where learners earn credits through real-world challenges, projects, pathway advisory, mentorship, and whole-human development.
Official FAQKhan World School
A 100% online grades 6-12 model with mastery-based progress, synchronous seminars, houses, squads, peer tutoring, and student-driven projects.
Official pageTUMO
Free teen centres built around self-learning activities, workshops, project labs, adaptive pathways, industry experts, and a living portfolio.
Education modelMinerva University
Small synchronous seminars, active discussion, interdisciplinary learning, global immersion, civic projects, and real stakeholders as part of the learning experience.
Official pageAlpha School
AI is used for diagnostic practice and personalised pacing, while human guides focus on motivation, habits, mentoring, and hands-on activity.
Official explainerThe chosen ARTIZAI build path
Build ARTIZAI as a hybrid virtual innovation school: Sora-style online studio culture, TUMO-style project pathways, Khan-style mastery evidence, School of Humanity-style real-world missions, Minerva-style live critique, and Alpha-style adaptive practice. The unique ARTIZAI layer is the seven-lab AI curriculum plus physical making kits, local hubs, British craft mentors, and a UCAS-ready portfolio.
Virtual school concept
ARTIZAI Cloud Studio: a virtual academy with local making moments.
ARTIZAI can run beyond one building. The virtual school lets learners join from any region, enter a Lab, work through AI-guided missions, meet live in studio seminars, publish portfolio evidence, and attend optional local make days through partner schools or city hubs.
AI Diagnostic Onboarding
Students map interests, subject confidence, tool readiness, safety needs, and preferred learning rhythm before joining a Lab.
Live Studio Seminars
Small groups meet weekly for critique, debate, planning, and showcase practice with a facilitator and peer cohort.
Mission Dashboard
Each sprint has research tasks, AI tool tasks, mentor checkpoints, build evidence, ethics reflections, and portfolio upload prompts.
Home Kits and Hub Days
Science, electronics, craft, and making kits support safe home builds; larger fabrication moves to school-hosted labs and city hubs.
Portfolio Passport
Every prompt, failure, iteration, mentor comment, image, code file, prototype test, and final reflection becomes evidence.
Credential Review
Students earn Bronze, Silver, Gold, or Pioneer credentials through project review, not passive attendance.
Build blueprint
Present ARTIZAI as an operating system for future learning.
The public story becomes simple: students join a Lab, run missions, build proof, publish a portfolio, and progress through credentials. The admin system tracks the work needed to turn that story into a launchable pilot.
School Identity
Mission, legal structure, launch region, brand voice, parent promise, partner promise, and safeguarding posture.
Learning Engine
Seven Labs, six age tiers, 12-week sprints, AI ethics thread, skill rubrics, and project bank.
Virtual Campus
Student dashboard, live seminars, AI toolbench, portfolio passport, parent view, mentor view, and facilitator console.
Physical Studio
Making kits, safe tool protocols, school-hosted labs, city hub scheduling, equipment inventory, and resident artisans.
Funding Stack
Innovate UK, UKSPF, DfE EdTech, Arts Council England, Wellcome, corporate sponsors, and bursary strategy.
Launch Machine
Pilot cohorts, school MOUs, admissions, mentor recruitment, evidence dashboard, showcase events, and expansion plan.
Use the admin side to complete the build steps.
The static admin dashboard saves your launch checklist locally in this browser and can export the plan as JSON for sharing or later development.
Impact case
Designed for students, families, schools, funders, and regional employers.
Roadmap
A phased UK rollout from North West pilot to national network.
-
P0 Foundation
Confirm entity, recruit education and regional directors, secure school and artisan support.
-
P1 Launch
North West pilot with Making & Engineering plus Science; first 30-50 student cohorts.
-
P2 Yorkshire Expansion
Add Mathematics and Arts labs, unlock cross-lab projects, and prepare first portfolio submissions.
-
P3 Midlands Full Secondary
Add Social Science and Languages, completing the core secondary offer across regions.
-
P4 Full Spectrum
Add Humanities, Seedlings, and Pioneers; grow toward 10+ operational centres.
-
P5 Franchise & EU Link
Open licensed delivery, hybrid platform support, and UK/EU student design challenges.
AQ
Project leadership
Built and led by a senior data-engineering founder.
Avais Ahmad Qarni Project CEO & Founder · CODES AI Limited
Lab specialists
Every lab is led by a domain expert who has actually shipped the work.
These profiles define the specialist standard for each ARTIZAI Lab.
Start the conversation
Build the first ARTIZAI pilot with schools, funders, and industry partners.
Use this starter site to open conversations with Innovate UK, DfE EdTech advisors, local authorities, schools, artisan partners, and advanced manufacturing sponsors.
North West first launch
Yorkshire and Midlands next
School, hub, and centre models