The accord is an explicitly voluntary commitment for companies to self-police on frontier AI vulnerabilities

The frameworks, policies, and standards that guide the responsible development, deployment, and oversight of artificial intelligence systems.