An AI ethical risk assessment addressing the ethical principles and potential harms is used when a project or programme begins and again whenever material changes to scope or outputs alter the risk profile, and it determines the approvals pathway. Senior leaders overseeing AI-enabled development and operation review the ethical risks and keep adequate, proportionate evidence of their risk management decisions, supported by suitably qualified and experienced personnel. Each AI use case receives an overall rating from impact and likelihood, based on residual risk, that fixes the approval level: critical to 2PUS or by exception Ministers, severe and major to Defence-level oversight such as the JROC/IAC, moderate and minor to TLB-level oversight with delegation; a programme may need JROC/IAC or ministerial approval even below JSP 892 thresholds.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.