← Back to Database
🏛 Governance Gap

UK AI Safety Institute Has No Enforcement Powers

✓ Verified
The UK AI Safety Institute (AISI) operates on a voluntary basis - it has no statutory authority to compel AI companies to submit models for testing or to enforce safety standards.
Verified: 1 March 2026 · Last updated: 1 March 2026 · Jurisdiction: UK

The AI Safety Institute (AISI), established following the 2023 Bletchley Park AI Safety Summit, conducts pre-deployment testing of frontier AI models. However, this testing is entirely voluntary - companies choose whether to participate.

AISI has no statutory powers to:

  • Compel companies to submit models for evaluation
  • Block the release of models deemed unsafe
  • Impose penalties for non-compliance with safety standards
  • Set binding safety thresholds

This means the UK’s primary AI safety body relies on goodwill from the very companies it is supposed to oversee.

AISIenforcementUK policy