AI Competency Frameworks Compared: Why Simplicity, Judgement, and Measurement Matter More Than Complexity

Author: Rob Williams Assessment

Core Proposition:

AI literacy is judgement capability for an AI-augmented world.

This white paper compares leading international AI competency frameworks — including those developed by UNESCO and Cedefop — with a unified psychometric model used across:

  • Schools (SchoolEntranceTests.com)
  • Organisations (Rob Williams Assessment)
  • Individuals (Mosaic AI Skills Framework)

The objective is not to critique global frameworks, but to identify where complexity, generalisation, and lack of measurement reduce practical effectiveness.


The Context: Rapid Growth of AI Competency Frameworks

Over the past three years, AI competency frameworks have emerged rapidly, particularly in education.

Two of the most influential are:

  • UNESCO AI Competency Framework for Students
  • Cedefop-aligned European adaptations

These frameworks aim to define what learners should understand about AI.

However, a critical question remains:

Do these frameworks measure capability — or simply describe it?


Overview of the UNESCO Framework

The UNESCO framework focuses on:

  • Human-centred AI understanding
  • Ethics and responsibility
  • Technical awareness
  • Societal impact

It is comprehensive and globally applicable.

Strengths

  • Strong ethical foundation
  • Global relevance
  • Policy alignment across countries
  • Broad conceptual coverage

Limitations

  • High conceptual complexity
  • Difficult to operationalise in classrooms
  • No direct measurement model
  • Limited linkage to workplace performance

Key issue:

The framework describes what students should understand, but not how capability is demonstrated or measured.


Overview of the Cedefop Interpretation

The Cedefop adaptation emphasises:

  • Vocational readiness
  • Digital competence integration
  • Skills frameworks alignment

Strengths

  • Closer link to employability
  • Structured competency mapping
  • Alignment with European skills policy

Limitations

  • Still largely descriptive
  • Complex multi-layered structure
  • Limited psychometric grounding
  • Difficult for practical implementation

Key issue:

The framework improves structure but still lacks measurement clarity.


The Core Problem with Existing Frameworks

Across both UNESCO and Cedefop approaches, a consistent issue emerges:

They define domains, not capability.

This leads to:

  • Difficulty in teaching
  • Difficulty in assessing
  • Difficulty in linking to outcomes

Most critically, they do not answer:

Can this individual make good decisions using AI?


The Alternative Model: A Simpler, Transferable Framework

The proposed model used across SET, RWA, and Mosaic is intentionally simpler.

It focuses on three core constructs:

  • Judgement
  • Reasoning
  • Problem solving

These are:

  • Psychometrically grounded
  • Measurable
  • Transferable across contexts

This simplicity is not a limitation.

It is the strength of the model.


Direct Comparison

CriteriaUNESCO / CedefopRWA / SET / Mosaic Model
StructureMulti-layered, complexSimple, three core constructs
FocusKnowledge and domainsCapability and decision-making
MeasurementLimitedCentral to design
UsabilityDifficult to implementHighly practical
TransferabilityPrimarily educationSchools, workplace, individuals
Psychometric groundingLimitedCore foundation

UK-Specific Relevance vs Global Universality

UNESCO frameworks prioritise global applicability.

This is valuable — but creates trade-offs.

Global Framework Strength

  • Consistency across countries
  • Policy alignment

Global Framework Limitation

  • Less tailored to UK education system
  • Less aligned with UK assessment culture
  • Less relevant to UK hiring practices

The RWA/SET/Mosaic model provides:

  • UK relevance
  • Alignment with assessment standards
  • Applicability to hiring and leadership

But remains globally transferable due to its simplicity.


Application Across Three Contexts

Schools (SET)

AI Literacy Skills Training

  • Develop judgement early
  • Focus on reasoning processes
  • Use scenario-based learning

Organisations (RWA)

AI Capability Diagnostics

  • Assess decision quality
  • Improve hiring validity
  • Reduce organisational risk

Individuals (Mosaic)

Mosaic AI Skills Framework

  • Measure capability
  • Track development
  • Identify gaps

Why Simplicity Wins

Complex frameworks often fail in practice.

Reasons include:

  • Difficult to teach
  • Difficult to assess
  • Difficult to apply consistently

A simpler model:

  • Improves adoption
  • Enables measurement
  • Supports consistency

Simplicity enables scale.


The Psychometric Advantage

The key differentiator is measurement.

The model supports:

  • Scenario-based assessment
  • Work sample simulation
  • Capability scoring

This enables:

  • Validity (prediction of outcomes)
  • Reliability (consistency)
  • Fairness (bias control)

This is largely absent in existing frameworks.


Graduate Employability: The Missing Link

Most frameworks stop at education.

This model extends into:

  • Early career capability
  • Hiring decisions
  • Leadership judgement

This creates a continuous capability pathway.

From classroom → to workplace → to leadership.


Conclusion: Defining the Next Generation of AI Frameworks

UNESCO and Cedefop frameworks provide valuable foundations.

However, they are:

  • Descriptive rather than measurable
  • Complex rather than practical
  • Educational rather than cross-domain

The alternative model offers:

  • Simplicity
  • Measurement
  • Transferability

AI literacy is judgement capability for an AI-augmented world.

This definition enables:

  • Better education outcomes
  • Improved hiring decisions
  • Stronger leadership capability

Next Steps

Book a consultation:

Schedule a call