Delphi findings

Responsibility

Who do experts think should bear responsibility for addressing AI risks?

Actor Responsibility Matrix

Expert consensus mapping responsibility for addressing each risk category across the AI ecosystem. Enables accountability frameworks and helps organizations understand their role in AI risk management.

Rows represent risks and columns represent actor types (e.g., developers, deployers, governments). Each cell shows the level of responsibility experts assign to that actor for managing the risk. Higher values indicate greater perceived responsibility, and checkmarks indicate areas of expert consensus. Scan across rows to see who is responsible for a given risk, or down columns to see where a given actor plays the largest role.

Actor Vulnerability vs Responsibility Butterfly

Compares the responsibility experts assign to each actor type against the vulnerability they attribute to it for the same risk.

Rows represent risks and columns represent actor types. In each cell, the red bar (left) shows responsibility experts assign to that actor; the blue bar (right) shows vulnerability. Longer bars mean stronger expert agreement. Compare the two within a cell to see whether responsibility and exposure align — and scan down a column to see where an actor's pattern of responsibility and vulnerability diverges across risks.

More themes

Explore other visualizations