AITOPAI热报
USGuidelineAdoptedModel providersDeployers

Superintelligence Agreement

超级智能协议
White House·Published 2026-09-29·Checked by 管理员 on 2026-10-06

On September 29, the White House and CEOs of 6 tech companies signed a Superintelligence Agreement, requiring AI companies to implement four safety measures: self-monitoring, team oversight, external audits, and independent committee management.

Source ↗Remind meUS · All

What it means for you

  • Companies must establish internal mechanisms to monitor AI model behavior
  • Need to form dedicated teams to ensure safety measures are effectively implemented
  • Must accept external independent audits and establish an independent committee on the board

    • What to do

    EngineeringEstablish a self-monitoring system for AI models
    ComplianceForm an AI safety oversight team
    LegalEngage third-party organizations for safety audits
    EveryoneEstablish an independent AI safety committee on the board

    Key dates

    2026-09-29Published

    Source documents

    Source ↗

    Coverage

    Stories in the feed
    Editorial summary, not legal advice; dates and status follow the original document. · Report a correction