Comment by Google DeepMind

A model for which the inherent risk assessment indicates a ML R&D CCL has been reached will be deemed to pose an acceptable level of residual risk for further development or deployment, if: We assess that the deployment mitigations have brought the residual risk of severe harm to an acceptable level, based on considerations such as the effectiveness of mitigations, and information pertaining to model propensities and the severity of related events. This is required only for external deployment and high-risk internal deployment, not further development.