The AI Governance Lifecycle

The AI governance tasks are mapped to the OECD’s AI system lifecycle framework.

Click here to return to the main page.

Citation: Mäntymäki, M., Minkkinen, M., Birkstedt, T., & Viljanen, M. (2022). Putting AI Ethics into Practice: The Hourglass Model of Organizational AI Governance (arXiv:2206.00335). arXiv. https://doi.org/10.48550/arXiv.2206.00335

 

Scroll down for an explanation of the AI Governance Lifecycle

The columns represent the lifecycle stages for
different AI governance tasks.

  • These include:
  • 1a) Planning and design
  • 1b) Data collection and preprocessign
  • 1c) Model building and interpretation
  • 2) Verification and validation
  • 3) Deployment
  • 4) Operation and monitoring

The rows represent different task categories.

The individual AI governance tasks from different categories are mapped along the lifecycle stages.

In total, there are 67 tasks.