AI Learner Safeguarding

An AI feature in a classroom can produce something harmful, or a learner can raise something that no software should be handling. This policy sets what happens next.

What it covers

• All live and asynchronous training where AI-generated or AI-supported output is visible to learners.
• Our current course set, which is designed for adult professional learners aged eighteen and over.
• Trainer responses to distress, welfare concerns, inappropriate AI output, harmful advice, boundary issues or disclosures during AI-supported learning.

What we commit to

• AI is not used to give therapeutic, clinical, safeguarding, legal or other professional advice outside the educational scope of the program.
• Where AI simulates conversation or offers suggestions, learners are reminded it is not a human professional and may be incomplete or wrong.
• There is always a clear human route during live delivery. The lead trainer and support trainer remain available whenever AI features are enabled.
• If a learner discloses or shows a welfare concern, the trainer responds calmly, avoids unnecessary probing, records only what is necessary and escalates through the designated safeguarding route and, where appropriate, the client's nominated contact.
• Immediate danger is escalated through the venue or client emergency process, or to emergency services. AI is never the decision-maker in an emergency.
• Safeguarding incidents linked to AI are recorded separately from ordinary content errors, with restricted access to the record.
• Any future program involving learners under eighteen requires a separate safeguarding review before delivery. This policy does not by itself establish child-safeguarding procedures.

The trainer boundary

A trainer is not expected to diagnose, counsel or investigate a welfare concern. The required response is to keep the learner safe, stay within their competence, protect privacy and escalate to the appropriate human route.