js
← Back to projects

AI Educator CPD

Designing practical professional learning to help primary teachers use generative AI critically, safely and confidently.

  • Learning design
  • UX design
  • Accessibility
  • Front-end development
  • Responsible AI
AI Educator module showing a ten-section learning journey, progress indicator and an explanation of generative AI

Generative AI created a difficult professional-development problem: primary teachers needed a useful starting point, but the subject was changing quickly and carried real risks around pupil data, inaccurate outputs and uncritical adoption.

I designed and built a self-directed digital CPD module that moves from basic orientation to realistic school scenarios, risk checks and a small plan for responsible experimentation.

The result is a working, responsive learning product. Its effectiveness with teachers still needs evaluation.

Role
Learning designer, UX designer & front-end developer
Audience
Primary teachers and school staff
Format
Self-directed interactive CPD
  1. 01DiscoverContext and constraints
  2. 02DefineDecisions and scope
  3. 03DevelopLearning experience and delivery
  4. 04OutcomeDelivered work and next validation
01Discover

The challenge

The challenge was not a shortage of AI information. Teachers were already encountering an excess of tools, claims and warnings. The more useful question was: what does a teacher need to understand and decide before using an AI-generated output in a real school context?

The experience needed to make AI understandable without oversimplifying it, connect abstract risks to familiar tasks and keep the teacher’s professional responsibility visible at every stage.

It also had to work as short, independent CPD without requiring a facilitator, user account or specialist learning platform, while avoiding the suggestion that AI was inherently beneficial or safe.

02Define

Three strategic decisions

I used three principles to keep the experience relevant, bounded and usable without a facilitator.

  1. Start with teacher tasks, not AI products

    Examples begin with planning, adapting resources and checking work so the technology remains connected to recognisable professional goals.

  2. Keep professional judgement visible

    Every scenario separates the situation, example prompt and teacher check. Outputs are positioned as material to review, not answers to accept.

  3. End with a bounded experiment

    The final plan asks teachers to identify one task, protected data, a checking method and a clear definition of success.

03Develop

A learning journey from orientation to action

The ten-section experience uses a persistent navigation list, progress feedback and short interactions to make a longer topic feel manageable. Learners move through explanation, classroom examples, a matching activity, risk guidance, a knowledge check and an action plan.

Progress and planning notes are stored locally in the browser. This keeps the experience lightweight and removes the need to transmit the teacher’s draft plan to an external service.

  1. Understand
  2. Explore
  3. Practise
  4. Check
  5. Plan
Classroom planning scenario with the situation, an example prompt and a teacher check presented as separate elements
A classroom scenario keeps the teaching situation, prompt and professional check distinct.

Accessibility and responsible use

Responsible use is taught as a decision process: which task, which data, which tool and which checks. Guardrails cover data protection, hallucinations and sustainability, with direct actions teachers can take before using an output.

The interface uses semantic controls, visible keyboard focus, labelled progress, readable hierarchy and responsive layouts. Content is divided into short explanations and concrete decisions to reduce cognitive load for people who may be unfamiliar with AI. These are implemented accessibility measures, not a claim of formal WCAG conformance.

AI use guardrails covering data protection, hallucinations and sustainability
Guardrails connect each risk to a concrete checking behaviour.
Mobile layout of the module with a visible focus outline around the Previous button
The same journey reflows for small screens and retains visible keyboard focus.

AI-assisted delivery with human review

I used AI to accelerate implementation while retaining control of the learning sequence, safeguarding boundaries, interaction choices and final review. Generated patterns were treated as drafts: I manually reviewed navigation, semantic controls, focus behaviour, local-storage messaging and responsive layouts before deployment.

  1. Research and content structure
  2. Interaction prototype
  3. AI-assisted implementation
  4. Manual QA and refinement
Four-field AI use plan for task, protected data, output checks and success, followed by a local-storage notice
The final activity saves a small action plan in the learner’s browser; no form data is sent elsewhere.
04Outcome

Delivered product, with impact still to test

The central lesson was that responsible AI education should begin with the task, the data and the professional judgement required, rather than the novelty of a tool. Making the teacher check visible throughout the journey helped keep that boundary clear.

Delivered

A complete learning experience

The ten connected sections combine interactive practice, progress tracking and a locally saved action plan. The working module demonstrates the intended structure and safeguards across desktop and mobile.

Next validation

Evaluation with teachers

Moderated testing should examine where teachers pause or skip, whether the scenarios transfer to their own work and whether the final plan feels specific enough to use. Screen-reader testing, organisational sign-off and telemetry remain future product work.