SECAI Core · phase 7 of 15
Input quota
The maximum amount of data or the maximum number of requests a user or application can send to an AI system within a time frame
The Explain card
- Plain English
- An input quota is the maximum amount of data or number of requests a user or application may send to an AI system within a time frame.
- Example
- An API key is limited to a few hundred requests per hour and a fixed token budget per day. When a compromised script starts hammering the endpoint at midnight, the quota cuts it off long before the bill or the model's patience runs out.
- Why it matters
- Quotas limit the damage from stolen keys, denial-of-wallet attacks, and automated probing used for model extraction. They also make abuse visible, since hitting the ceiling is a signal worth alerting on.
- Hook
- An input quota is a speed limit on the model's front door.
Where it sits in the deck
Phase 7: AI-Specific Runtime Controls: Guardrails, Limits, and Enforcement
Responsible AI principles must be enforced mechanically — learn the runtime controls that constrain model inputs, outputs, and resource consumption.
