The ChalkboardAI Approach

At the heart of ChalkboardAI is a simple belief: teachers deserve AI tools that respect privacy, reduce workload, and reflect real classroom practice.

The ChalkboardAI Approach is built around six principles:

• Local-first data handling - identifiable pupil data stays on the teacher’s device.
• Automatic anonymisation - names are replaced with pseudonyms before anything touches the model.
• Contextual minimisation - each tool uses only the information required for the task.
• Local reidentification - real names never reach the model; outputs are reidentified safely on-device.
• Human-in-the-loop - teachers remain the final decision-makers.
• Transparent reproducibility - every operation reveals the anonymised prompt and metadata used.

This approach creates an ecosystem where AI is safe by default, practical by design, and genuinely aligned with the way teachers work.

The full white paper outlines this architecture in detail and sets out a roadmap for responsible AI use in schools.