GenAI framework, Incident Reporting: set up channels for outsiders to report safety vulnerabilities, with a window to assess, patch and publish as in software vulnerability practice, plus ongoing monitoring to catch malfunctions; after incidents, have internal processes for timely notification and remediation, including to the public or government depending on impact, define severe AI incidents or a materiality threshold for formal reporting, and align with existing reporting regimes.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.