Our approach
ECHO uses AI to help authorised safeguarding staff organise and prioritise reports, not to replace professional judgement. People remain responsible for reviewing reports and deciding what action to take. We minimise the information collected, protect it with appropriate security controls, and clearly explain how data is used, stored and shared.
AI supports triage by helping authorised staff organise information and identify reports that may need prompt attention. Safeguarding professionals remain responsible for reviewing every report, assessing context, and making safeguarding or disciplinary decisions.
ECHO is designed to collect only the information needed to receive, assess and follow up a report. Access should be limited to authorised people, supported by clear retention controls and appropriate technical and organisational safeguards.
Report information is not used to train general-purpose AI models unless that use has been explicitly and lawfully agreed. How information is processed by service providers should be documented and kept under review.
AI-supported outputs can be incomplete or wrong. Performance should be monitored for errors and bias, important decisions should be recorded, and people should be able to question or challenge an outcome.
These commitments align with New Zealand public-sector guidance emphasising transparency, privacy, accountability and meaningful human oversight when using generative AI.
