Boundary design
Designing where an AI system's authority starts and stops: which actions it may take, with what data and permissions, and where it must hand back to a person or recover.
Boundary design is the work of setting limits around an AI system so it can be useful without being dangerous. It covers which actions the system may perform, on which data, under which identities and permissions, with what monitoring, and where it must stop and return a case to a person.
It matters now because the central constraint has moved. In the phase of capability discovery, the question was what models could do. As capability becomes widely available, the scarce thing is a well-drawn boundary: authority that is as wide as is worthwhile, and no wider.
Elements of a boundary
- The maximum authority to delegate while failures remain detectable, containable, reversible and economically worthwhile
- A threshold policy for when to act, defer or escalate
- A bounded blast radius
- A path for recovery
The sharpest AI questions are therefore about boundaries, not intelligence.
Read more in The wrong questions about AI right now.
Related terms
Capability discovery
The first phase of generative AI, when attention went to finding out what models could do. The next phase, boundary design, is about where they should act and with what authority.
Capability is not authority
A design principle for AI agents: being technically able to perform an action does not mean the agent should be permitted to perform it. Delegation needs gradients of authority.
Blast radius
How far the damage of a failure can spread: which data, systems, people and money an AI action can affect if it goes wrong. Smaller radius means safer delegation.
Threshold policy
The explicit rules that turn a model's probability or score into an action — reject, send to a human, or act autonomously — owned and reviewed separately from the model itself.