Definitions of AI literacy

Broadly speaking, the term AI literacy is used to describe "a core set of knowledge, skills, and values for interacting ethically and effectively with AI in the present" (UNESCO, 2024, p.11).

It may be conceptualised as comprising four key aspects (Ng et al., 2021):

Know and understand AI
Use and apply AI
Evaluate and create AI
AI ethics

AI literacy, or AI competence, is more precisely described through frameworks, which may be used to incorporate it into educational curricula (see below).

AI literacy as digital competencies

In the European Union, there is currently no specific framework dedicated to defining AI literacy for students. However, the Digital Competence Framework for Citizens (DigComp) has been updated to include elements related to AI in version 2.2. These updates show how AI is currently thought to interrelate with the development of digital competencies, especially information and data literacy, in EU citizens:

Competency Area	Description
Browsing, searching, and filtering	Effectively querying Generative AI tools must become part of citizens’ digital skills. At the same time, citizens must be aware of the challenges in filtering AI-generated output.
Evaluating information	Citizens must become aware of potential biases in AI datasets, the risk of algorithms creating "echo chambers," as well as AI's ability to skew information, misrepresent meanings, and produce "deepfake" content.

Figure: European Union’s digital competence framework (DigiComp 2.2). Click on the image to enlarge.

DigiComp 2.2

AI literacy according to UNESCO

At present, the UNESCO AI competency framework for students represents the most comprehensive framework that defines AI literacy for higher-education students and supports its integration into curricula. The framework emphasises the importance of developing not only AI-related knowledge and technical skills but also, and especially, critical thinking skills and value judgments.

The framework (see figure below) comprises four competencies and three progression levels, with the final competency specifically targeting AI developers. This design allows for the gradual integration of AI literacy into the curriculum and enables differentiation in the technical skills taught based on the specific study program. The four competencies are summarised below. Refer to the original publication for further details and examples of educational methods for each competency and level:

Competency	Description
Human-centred mindset	The values, beliefs, and critical thinking skills needed for students to examine whether the AI system is sound, its use justified in specified contexts, how humans should interact with it, and how to foster safe, inclusive, and just AI societies.
Ethics of AI	The ethical judgment, reflective skills, and social-emotional competencies needed to navigate and adapt to evolving principles and regulations governing AI throughout its lifecycle.
AI techniques and applications	The foundation in technical knowledge and operational skills that allow students to gain insight into how AI is developed and make informed decisions about its use and development.
SI system design	This competency relates to the application of the systemic design thinking methodology to the design, development, testing, and refinement of AI solutions and is intended only for students who wish to specialise in the development of AI.

Figure: UNESCO’s AI competency framework for students (p.19). Click on the image to enlarge.

AI literacy at UT

A working group at CELT led by Marcus Pereira Pessoa, Associate Professor at the ET Faculty and CELT’s Head of Educational Innovation, is currently working on a framework for AI literacy at UT.

✉️ Do you have questions or would you like to give input to the development of a UT AI literacy framework? Contact m.v.pereirapessoa@utwente.nl