US and EU Sign Historic AI Safety Treaty Setting Global Standards

The United States and European Union have signed the first international AI safety treaty, establishing shared standards for testing, deployment, and oversight of advanced AI systems that both regions will enforce.

Key Provisions

The treaty mandates safety evaluations for frontier AI models, creates a joint US-EU AI Safety Board, requires incident reporting within 72 hours, and bans certain AI applications including autonomous weapons and mass surveillance.

Global Implications

China, Japan, and the UK have indicated interest in joining the framework, which could become the global standard for AI governance.

• First international AI safety treaty
• Joint US-EU AI Safety Board
• Mandatory safety evaluations for frontier models
• Bans autonomous weapons and mass surveillance