CreaComp

Competency 4.3

Create with artificial intelligence

Boundary this domain describes capacities of judgement and direction, never software. No tool is named in its descriptors: tools change every six months, the capacities that govern them remain.

Definition

Directing generative systems the way an art director directs a supplier: precise, iterative instructions, qualitative judgement of outputs, reasoned choice of what is delegated and what stays human. The decisive competency is not knowing how to produce with AI — everyone can — but knowing how to judge and direct what it produces.

Components

  • Formulating instructions: context, constraints, examples, success criteria
  • Iteration: critiquing an output, refining, rephrasing until the target level is reached
  • Qualitative judgement: telling a mediocre output with no visible error from a genuinely good one
  • The human/machine split: deciding what to delegate, what to keep, what to verify
  • Advanced techniques: decomposition into steps, worked examples, chained instructions
  • Traceability: documenting your choices in order to own them and pass them on

The four levels

  1. N1 Discovery Lower secondary · complete beginner

    Is able to obtain a usable result from an AI by specifying context and constraints, and to correct an inadequate output.

    Evidence of assessment

    Present the before/after of an improved instruction, explaining the changes and their effects.

  2. N2 Application Upper secondary · advanced beginner

    Is able to tell a good output from a mediocre one with no visible error, and to rework it up to the target level.

    Evidence of assessment

    Rank and justify three AI outputs of differing quality, then deliver an edited final version.

  3. N3 Adaptation Tertiary · junior professional

    Is able to produce a complete piece of work by steering the AI and to justify each split between human and machine.

    Evidence of assessment

    Produce a complete piece of content documenting every decision — made by hand or delegated to the AI — and justifying the choices.

  4. N4 Orchestration Experienced professional

    Is able to design reusable instruction systems and to architect a production chain where AI occupies a decided, controlled and revisable place.

    Evidence of assessment

    Design a system of five chained instructions producing a piece of content end to end, documented, with its human checkpoints and what the model was refused.