Eight domains of competence
D1 and D2 are foundational to all others. D3 and D4 are the domains of doing. D5 and D6 are the domains of judging. D7 and D8 constrain all six.
Published 8 of 8 domains · version 1.29
| Ref | Domain | Scope | Statements | Status |
|---|---|---|---|---|
| D1 | AI foundations | What AI systems are, how they behave, and what they cannot do. | 22 | PUBLISHED |
| D2 | Data and information | Sourcing, judging and governing the information AI systems consume and produce. | 23 | PUBLISHED |
| D3 | Interaction and elicitation | Obtaining reliable results through deliberate interaction. | 19 | PUBLISHED |
| D4 | Automation and integration | Connecting AI systems to others, and running them unattended. | 22 | PUBLISHED |
| D5 | Solution design | Deciding what to build, whether to build it, and how it fits together. | 22 | PUBLISHED |
| D6 | Evaluation and assurance | Establishing with evidence whether a system is good enough, and when it stops being. | 22 · L1 4 · L2 6 · L3 8 · L4 4 | PUBLISHED |
| D7 | Ethics, governance and professional practice | Obligations to other people arising from using AI at work. | 22 | PUBLISHED |
| D8 | Security | AI systems under adversarial conditions, and containing their behaviour. | 25 · L1 5 · L2 7 · L3 9 · L4 4 | PUBLISHED |
Each domain page states its full scope, exclusions, boundary notes and editorial position.
Drafting sequence
One domain at a time, in dependency order
Each domain is drafted, reviewed and confirmed before the next begins. Drafting several in parallel produces level drift that is expensive to correct once identifiers are published and cited.