Towards Superintelligence Alignment
agendareviewed

GovAI / UK AISI (governance & eval)

Can governance mechanisms and institute evaluations keep pace with capability and actually bind deployment decisions under race pressure?

Introduction

GovAI and allied institutes connect policy research to frontier evaluation, touching Goodhart Selection, Inner Alignment, and Deployment Safety in institutional form.

Who carries it: Centre for the Governance of AI (Oxford); UK AI Security Institute; US CAISI (NIST); Simon Institute; IAPS (policy orbit)

What they aim to do. Inform decision-makers and test frontier systems through policy research and government-led evaluation capacity.

The hard question. Can governance mechanisms and institute evaluations keep pace with capability and actually bind deployment decisions under race pressure?

What they produce. Policy-facing research, the GovAI Fellowship, gradual disempowerment programmes, frontier model testing, standards development, and alignment evaluation case studies.

Key terms. Terms include AI governance, compute governance, frontier evaluations, gradual disempowerment, AI safety institutes, pre-deployment testing, and evaluation awareness.

Related field cruxes. Goodhart Selection; Inner Alignment; Grounding Drift; Deployment Safety

What they contribute. A governance research pipeline; EU and US policy translation; gradual disempowerment framing; institutional evaluation capacity; Resolution and Irving lineage via UK AISI; personnel and advisory overlap between GovAI and AISI.

How this project treats it. Policy levers and evaluation visibility do not imply correction-channel integrity .

Map clustering

AISafety.com map listings that roll up to this agenda:

See the coverage matrix for evidence tagged to this agenda, and the glossary for shared terms.