School Entrance Tests AI Competency Model for Schools ,

Welcome to our AI competency model for schools. This is based upon the universal MosAIc AI Skills Framework. It is designed by Rob Williams, a highly experienced psychometrician. Hence, it is ideal to use as the basis for our own AI psychometric assessments and for schools to use as the basis for their own AI psychometric assessments.

AI Competency Model for Schools

There are eight AI competencies in our model:

Understanding AI competency
Prompting AI competency
Evaluation AI competency
Decision-making AI competency
Ethical awareness AI competency
Workflow use AI competency
Credibility judgement AI competency
Confidence AI competency

There are aligned AI competency models for organisations and AI competency models for individuals. All are based upon the MosAIc universal AI skills framework.

Higher level AI skills

However, a single-layer model is insufficient to capture AI capability. Two layers are needed:

A skills framework to explain underlying ability but not performance
And a competency model alone measures behaviour but not its causes

The combination provides a complete system:

Mosaic pillars → underlying capability structure
AI competency framework → observable performance

This distinction mirrors established psychometric practice:

latent traits vs behavioural indicators
constructs vs outcomes

The universal MosAIc AI Skills framework

The universal MosAIc AI skills framework defines the underlying skill structure that supports AI capability. It identifies nine core skills:

Analytical Reasoning skills
Cognitive Flexibility skills
Ethical Judgement skills
Information Credibility skills
AI Output Validation skills
Structured Decision-Making skills
Bias Recognition skills
Learning Agility skills
Attention Control skills

These pillars describe the foundational abilities that influence how individuals engage with AI systems. However, while this model explains why individuals differ in their performance, it does not directly measure how well they perform in practice. To address this, a complementary model is required.

Understanding Each AI Competency in the MosAIc AI Competency Model

1. Understanding AI competency

This capability refers to an individual’s grasp of how AI systems generate outputs. It includes:

recognising probabilistic generation
understanding limitations of training data
awareness of hallucination risk

Critically, it is not technical depth that matters, but functional understanding. An individual does not need to build a model, but must understand enough to interpret its outputs appropriately.

2. AI Prompting competency

Prompting is often mischaracterised as a technical skill. In practice, it is a function of:

clarity of thinking
ability to structure information
iterative reasoning

Effective prompting involves:

specifying constraints
refining inputs
recognising when outputs require adjustment

It is less about “knowing tricks” and more about cognitive flexibility and precision.

3. AI Evaluation competency

Evaluation is the capacity to assess the quality of AI-generated outputs. This includes:

identifying inaccuracies
detecting omissions
assessing relevance

It is one of the most critical capabilities, and one of the least developed. A key risk in AI use is false confidence in plausible outputs. Evaluation mitigates this risk.

4. AI Decision-Making competency

AI systems do not make decisions. They provide inputs into decisions. This capability involves:

integrating AI outputs with other information
weighing uncertainty
making informed judgements

It is particularly important in high-stakes contexts, where over-reliance on AI can lead to significant consequences.

5. AI Ethical Awareness competency

AI use raises ethical considerations across multiple domains:

fairness
bias
accountability
transparency

Ethical awareness is the ability to:

recognise these issues
anticipate potential risks
act responsibly

It is not a compliance exercise, but a judgement capability.

6. AI Workflow Use competency

This capability reflects how effectively individuals integrate AI into their work processes. It includes:

knowing when AI adds value
avoiding unnecessary use
combining AI with human judgement

Effective workflow use is characterised by selective and purposeful application.

7. AI Credibility Judgement competency

Credibility judgement is the ability to determine whether an AI output should be trusted. It involves:

assessing source reliability
recognising signals of uncertainty
identifying when verification is required

This capability is central to managing risk in AI-assisted environments.

8. AI Confidence competency

Confidence reflects how individuals perceive their own capability. It influences:

willingness to use AI
susceptibility to over-reliance
openness to revision

Importantly, confidence must be calibrated:

overconfidence leads to error
underconfidence limits effectiveness

How the Competencies Interact

These eight AI capabilities do not operate in isolation. For example:

Prompting influences evaluation
Understanding AI shapes credibility judgement
Evaluation informs decision-making
Ethical awareness constrains decisions

This creates a system of interdependent processes. Weakness in one area can undermine performance across others.

Mapping MosAIc Competencies to MosAIc Skills

The relationship between the two models can be understood as follows:

Analytical Reasoning → Evaluation, Decision-making
Information Credibility → Credibility judgement
Cognitive Flexibility → Prompting, Workflow use
Ethical Judgement → Ethical awareness
Bias Recognition → Evaluation, Credibility judgement
Attention Control → Prompting, Workflow use
Learning Agility → Understanding AI, Workflow use
AI Output Validation → Evaluation
Structured Decision-Making → Decision-making

This mapping demonstrates how underlying skills translate into observable behaviours.

Implications for Education

In educational contexts, AI is often introduced as a tool to be learned. This approach risks:

superficial engagement
over-reliance
limited transferability

A competency-based model shifts the focus to:

developing judgement
strengthening evaluation skills
building responsible use

This is more aligned with long-term capability development.

AI Literacy for Students

AI tools are now part of everyday school life. Students are using them for homework, revision, and coursework. But a critical question is often missed: Does your child know how to use AI well, or just how to use it?

What AI Literacy Really Means for Students

AI literacy is not about typing questions into ChatGPT. It is about:

Understanding when AI can be wrong
Checking answers properly
Using AI safely and responsibly
Thinking independently rather than relying on AI

For background, see this explanation of AI.

Why Schools Are Concerned

Reports in The Guardian highlight growing concerns about AI use in schools. The risks include:

Over-reliance on AI
Reduced critical thinking
Plagiarism and academic integrity issues
Exposure to incorrect information

The AI Skills Competency Framework for Students

We assess student AI literacy across key areas:

Understanding AI limitations
Evaluating answers
Using AI safely
Making independent decisions

How Parents Can Support AI Literacy

Encourage questioning of AI answers
Discuss when AI should not be used
Focus on understanding, not shortcuts

From School to Future Skills

AI literacy is now a core future skill. For advanced capability development, explore AI skills for individuals or organisational readiness via AI readiness assessment designers Rob Williams Assessment Ltd.

Why AI Skills Matter for Careers

AI literacy is becoming a core employability skill. Those who can:

Use AI effectively
Evaluate outputs critically
Apply AI in workflows

will outperform those who rely on AI passively.

FAQ

Is AI good for learning?

Yes, when used properly. The key is developing strong evaluation and thinking skills.

What is AI literacy for students?

The ability to use AI safely, critically, and effectively in learning.