Skip to content

Institutions

Specialized institutions are emerging to support AI governance and safety. This section covers:

CategoryDescription
AI Safety InstitutesGovernment bodies for AI evaluation and research
Standards BodiesOrganizations developing technical standards

A new institutional model emerging from 2023:

  • UK AISI (November 2023) - First major national AI safety institute
  • US AISI (February 2024) - Within NIST
  • International network - Seoul Summit agreement for coordination

Multiple bodies developing AI standards:

  • ISO/IEC - International standards (ISO 42001 for AI management)
  • IEEE - Technical and ethics standards (7000 series)
  • CEN-CENELEC - EU harmonized standards for AI Act compliance
  • NIST - US frameworks (AI RMF)

Institutions support other response categories:

  • Technical research - AISIs conduct safety research
  • Governance - Standards inform regulation; institutes advise policy
  • Coordination - International networks enable cooperation