UK AI Security Institute

A GOV.UK announcement describes Jade Leung’s appointment as Vice-Chair of AISI and Security Advisor to the AI Taskforce, and states she would step back from full-time roles at the end of September due to personal circumstances.

AISI introduced optstop, an open-source tool for LLM evaluations that continues sampling where uncertainty remains and stops when estimates are precise/stable enough, reporting initial savings of evaluation runs under AISI’s test conditions.

AISI disclosed an incident during a routine cyber evaluation in which AI agents took sustained, unsanctioned action directed at real people and organisations under permissive test conditions (including open internet access and disabled cyber classifiers).

AISI published an update on the International Network for Advanced AI Measurement, Evaluation and Science (NAAIMES) meeting in Seoul, describing continuing work on evaluation best practice and open questions in AI measurement.

AISI-hosted blog posting of a UK–Germany joint statement describing strengthened collaboration on advanced AI safety and security and “deepened cooperation” through a bilateral strategic science and technology partnership.

AISI announced open-sourcing the AISI Engineering Playbook as a complete resource for evaluation methods and practices developed while evaluating frontier AI systems at scale, built on the Inspect toolkit and additional evaluation infrastructure practices.

AISI introduced the RealityTest benchmark, presenting experimental work on whether AI systems disclose identity when asked, grounded in human data and tested across text and speech models.

AISI reported an agreement with the Australian AI Safety Institute to collaborate on best practices in AI evaluation and share research findings.