Back
conceptUpdated Apr 18, 2026
AI Lifecycle
ai-developmentai-governance
- Jurisdiction
- US-Federal
- Issuer
- NIST
The AI lifecycle encompasses all stages of an AI system's development and deployment, organized around key dimensions:
Lifecycle Stages:
- Plan and Design
- Collect and Process (Data)
- Build and Use (Model)
- Verify and Validate
- Deploy
- Operate and Monitor
Key Dimensions:
- Application Context: The setting and purpose for which the AI system is designed
- Data and Input: Information used to train and operate the system
- AI Model: The algorithms and computational components
- Task and Output: What the system does and produces
- People & Planet: Human rights and broader societal well-being considerations
The lifecycle involves diverse AI actors with different roles and responsibilities. Risk management should ideally begin with the Plan and Design phase and continue throughout all stages. The interdependent nature of lifecycle activities means that decisions in one stage can significantly impact others.
TEVV (test, evaluation, verification, and validation) processes should be integrated throughout the lifecycle, not just at specific checkpoints. The NIST AI Risk Management Framework functions can be applied at different lifecycle stages as appropriate.
Neighborhood