K Knowledge HF-D7.7-001 transversal 9 skills
AI literacy
Knows how AI systems work, what they can and cannot do, states where they fit or not and uses correct terminology.
How hrmforce measures this
- Assessment method
- Knowledge test · rho 0.40 (SD 0.13)
- hrmforce instrument
- Knowledge test (client-specific)
- Competency (50-framework)
- Learning Ability
- Trainability
- high
- Demand outlook 2026 to 2030
- rising
Test of declarative job knowledge, usually assembled per client.
Behavioural anchors
| Level | Behaviour at this level |
|---|---|
| N1 Guided | Explains in own words that a language model predicts from patterns and can therefore produce incorrect output. works under supervision and follows instruction · routine, one variable at a time · own task |
| N3 Proficient | Judges per work task whether AI is suitable, names the main limitations and explains how it works to colleagues. sets own approach and seeks input proactively · several variables, some ambiguity · own team or process |
| N5 Leading | Defines what AI literacy means in the organisation, measures the level and links it to training and deployment policy. sets the standard and the policy · strategic, under high uncertainty · organisation, value chain or profession |
N2 and N4 are deliberately not anchored. Raters place them between the anchors, following the O*NET convention.
Underlying skills
These skills inherit the assessment route and the behavioural anchors of this construct.
| T | Skill | Definition | Demand outlook 2026 to 2030 |
|---|---|---|---|
| K | Explaining how a language model works Large language model · LLM | Explains in plain language that a language model predicts the next word based on patterns. | rising |
| K | Understanding tokens and context window Tokens · Context window | Explains how text is split into tokens and why a model only takes a limited amount of context. | rising |
| K | Distinguishing forms of AI Types of AI | Distinguishes rule based systems, predictive models and generative AI and states what each is suited for. | rising |
| K | Understanding the origin of training data Training data | Explains that model output reflects the properties and gaps of the training data and what that means for use. | rising |
| K | Naming the limits of AI Limits of AI | Names where an AI system is unreliable or inappropriate, such as legal consequences, rare cases and current facts. | rising |
| K | Using AI terminology correctly AI terminology | Uses terms such as model, prompt, agent, fine tuning and hallucination correctly in conversation. | rising |
| V | Justifying a model choice Model selection | Chooses between models in a justified way on quality, speed, cost, language and data location and records the trade off. | rising |
| V | Controlling token costs Token cost management | Estimates and limits AI usage cost by managing prompt length, model choice, caching and number of calls. | rising |
| K | Knowing multimodal AI Multimodal AI | Knows that models can also process image, speech and documents and states where that is useful at work. | rising |