Human oversight
AI can support judgement, but humans remain responsible for decisions and their consequences.
Responsible artificial intelligence
We govern artificial intelligence across its full lifecycle so that educational purpose, safeguarding, fairness, privacy, accessibility and community benefit remain central.
Non-negotiable principles
AI can support judgement, but humans remain responsible for decisions and their consequences.
Potential harm is assessed and controlled before use, with enhanced scrutiny for children and vulnerable people.
Bias, accessibility, SEND impact and unequal outcomes are considered and monitored.
People should understand when AI is used, its purpose, its important limits and how to seek human review.
Define the educational or community need
Classify risk and consider non-AI alternatives
Assess data, safeguarding, accessibility and ethics
Test accuracy, bias, safety and educational suitability
Approve at a level proportionate to risk
Monitor performance, feedback, incidents and change
Retire safely and preserve appropriate records
We welcome questions about how AI is used in our programmes. Concerns, unexpected behaviour or potential harm should be reported so that use can be reviewed, paused or changed.
Ask about responsible AIPublic summary based on the approved Responsible AI Governance Framework (TAH-0241).