Theoretical Framework
Key Definitions
The definitions and terminology employed in AICET rely on those defined in international standardisation reference frameworks relating to competencies and Artificial Intelligence.
They notably incorporate concepts, vocabulary, and definitions from standards ISO/IEC 22989:2022 (Information technology — Artificial intelligence — Artificial intelligence concepts and terminology) and ISO/IEC 23053:2022 (Framework for Artificial Intelligence (AI) Systems Using Machine Learning (ML)), which establish the descriptive and functional foundations of AI systems.
These references ensure that AICET terminology is consistent with recognised practices and compliant with international standards, facilitating its interoperability and readability in technical, regulatory, and organisational contexts.
Theoretical Frameworks for Expertise Levels
The AICET methodology combines two reference frameworks to structure the AI domain and evaluate competencies:
Bloom-Krathwohl Taxonomy: Used to describe knowledge progression, transformed into a matrix comprising types of knowledge (factual, conceptual, procedural, metacognitive) and cognitive processes (remember, understand, apply, analyse, evaluate, create).
DigComp Framework: Offers a methodology to describe competency management in the digital domain.
The Three AICET Expertise Levels
Acculturation Level
Target Audience: People who have general competence in all aspects of AI, but are not necessarily very familiar with specific details. This level constitutes an entry point for everyone.
Evaluation Objective: Evaluate the ability to explain basic AI concepts, detect AI contribution in everyday tools, have ideas about its dangers and contributions in civil society, or be able to distinguish what is AI from what is not. The goal is to measure general understanding of AI.
Example of Competence: Be able to know and briefly explain the difference between supervised learning and unsupervised learning.
Advanced Level
Target Audience: Individuals with a deeper understanding of AI and who may even be regular users of AI-based applications. An academic knowledge equivalent (postgraduate diploma, Bachelor’s or Master’s in computer science and particularly in AI) is recommended.
Evaluation Objective: Evaluate understanding of basic AI principles and methods. This audience will be tasked with finding the type of AI to use to solve a particular problem. This level concerns mastering the construction of AI systems or putting them into operation.
Example of Competence: Be able to select the appropriate type of learning for a given problem, taking into account data nature and final objective, and understand the implementation of typical algorithms for each.
Expert Level
Target Audience: Individuals possessing deep technical knowledge in AI, such as engineers, researchers, or professionals working in the field. Extensive academic knowledge is recommended (equivalent postgraduate diploma, Master’s or PhD in computer science and particularly in AI).
Evaluation Objective: Test expertise and ability to explain complex and advanced concepts. Computer code corrections may be part of the test. This level requires extensive knowledge and an ability to innovate.
Example of Competence: Be able to solve complex problems thanks to a deep understanding of all types of machine learning.
A collaborative initiative supported by leaders in technology, academia, and standardisation
AICET was developed by an open working group of individuals and organisations committed to establishing a trusted benchmark for AI skills. This diverse coalition of experts ensures the standard is comprehensive, practical, and aligned with real industry needs.
Key Contributors
Standardisation and Coordination Partners
The standard is developed under the coordination of AFNOR and aligned with the European CEN-CENELEC framework, guaranteeing a rigorous and internationally recognised process.