A.4.2 Resource documentation
Identifying and documenting the resources each AI system life cycle stage needs: the umbrella record that ties together data, tools, compute and people, and how it supports an AI system inventory.
A.4 Resources for AI systems
The five Annex A controls for documenting the resources an AI system depends on: an overall resource record plus data, tooling, system and computing, and human resources.
A.4.3 Data resources
Documenting the data resources an AI system uses as part of its resource record: what is recorded about each dataset and how this control relates to the detailed data controls in A.7.
A.4.4 Tooling resources
Documenting the tooling used to build and run AI systems, such as model frameworks, pre-trained models and evaluation tools, and keeping that record current as tools change.
A.4.5 System and computing resources
Documenting the system and computing resources an AI system relies on, from training and inference infrastructure to hosting, and how capacity and environmental footprint are recorded.
A.4.6 Human resources
Documenting the people and competences involved across the AI system life cycle, from development and deployment to maintenance and decommissioning, and how competence evidence links to clause 7.2.
