arXiv:2611.24987v1 Announce Type: new Abstract: AI evaluation can involve an active, adaptive optimisation process in which AI agents re-plan, retry, use tools or delegate work to achieve a goal. Unanticipated capabilities or unsafe behaviour may emerge, and existing controls may be inadequate to prevent harm. Selected safeguards may also be deliberately relaxed to observe behaviour they would otherwise suppress, creating additional risks. This article examines the distinctive safety considerations of evaluation and their implications for control placement, observation and intervention, and presents four architectural patterns to support safe AI evaluation, including diverse supervisors, an independent verifier, silent monitoring and coordinated shutdown.
Continue to arXiv Software Engineering →You are being forwarded in 6s — stay on this page