SHORT DEFINITION
The organizational belief that someone else owns accountability for an AI agent's behavior — a belief that is usually held simultaneously by every team involved.
The Accountability Assumption is the organizational belief that someone else owns accountability for an AI agent's behavior — a belief that is typically held simultaneously by every team involved. IT assumes the business owns what business users build. The business assumes IT governs what touches enterprise data. Compliance assumes security owns the agent risk surface. Security assumes compliance owns the regulatory exposure. When an agent causes a governance failure, the Accountability Assumption is why every team can point at another team with equal confidence. It is not a coordination failure. It is a structural gap that low-code agent tools have made routine.
CANONICAL EXAMPLE
The OCC examiner asks who authorized the Copilot Studio agent that missed the Form ADV filing deadline. The compliance officer says IT owns agents. IT says the business owns what it builds in low-code tools. The CISO says neither team reported it. All three answers are simultaneously sincere and insufficient. That is the Accountability Assumption in operation.
USAGE GUIDANCE FOR CONTENT
Use The Accountability Assumption as the central tension in story-format posts and newsletter scenarios. It is most effective when a named fictional character — a CISO, a compliance officer, an examiner — surfaces the gap through a specific conversation or audit finding. Never use it abstractly. It lands when it is dramatized through a specific moment where multiple teams each believed someone else was responsible.