A Pedagogical Framework to Teach Artificial Intelligence from an

Uruguayan Experience

María Eugenia Curi

, Germán Capdehourat

, Brian Lorenzo, Emiliano Pereiro

and Víctor Koleszar

Ceibal, Uruguay

Keywords: Artificial Intelligence, Computational Thinking, Education, K-12, Competencies.

Abstract: This article presents a pedagogical framework for integrating Artificial Intelligence (AI) education into the

school curricula in Uruguay. Its aim is the development of students' AI literacy, focusing on critical thinking,

problem-solving, creativity, and collaboration. Emphasizing computational thinking, it prepares students for

the digital era and addresses the ethical and social implications of AI. The framework poses three key

questions: What is AI? How does it work? What can it do? From these questions, five dimensions are

established to cover the fundamentals of AI: the main definition of AI, the representation of knowledge,

machine learning, the computational approach and the ethical use and social impact involved. Additionally,

the framework outlines key principles and a list of competencies to guide educators and educational leaders

in creating an informed, responsible, and adaptive approach towards AI and its societal impact. This

comprehensive guide is instrumental for educators to effectively incorporate AI concepts and develop AI

skills in students.

1 INTRODUCTION

This framework aims to establish a structure for

designing study programs and planning of teaching

and learning activities that promote a deep

understanding of Artificial Intelligence (AI) and

competencies for students to analyze, design and

solve problems using computational principles.

Through an integrated and multidisciplinary

approach, it aims to develop AI literacy and enhances

the skills of critical thinking, problem solving,

creativity and collaboration. In addition, ethical and

social aspects associated with the use of AI will be

addressed, an approach that aims to promote an

informed and responsible reflection on its impact on

both society and individuals. By establishing a

contextualized competency framework, this

document aims to provide guidance to the community

of educators and education leaders on the integration

of AI and Computational Thinking (CT) within the

https://orcid.org/0009-0003-9610-2772

https://orcid.org/0000-0002-3975-2168

https://orcid.org/0000-0001-6622-7732

https://orcid.org/0000-0001-6666-6786

https://ceibal.edu.uy/en/what-is-ceibal/

school environment. The implementation of

innovative teaching strategies combined with the use

of appropriate technological tools are expected to

enhance the development of key skills and

competencies for the 21st century. The aim is also to

foster an open and adaptive mindset in each student,

in order to build foundations to face the challenges

and opportunities that AI and digital technology

bring.

In summary, this competency framework offers a

general approach to AI education to create critical and

ethical citizenship in the use and understanding of this

technology and its transformative potential.

This work is organized into five main sections to

provide a comprehensive exploration of the topic.

Section 2 of this article examines Uruguay's process

of integrating computational thinking and artificial

intelligence into primary education from Ceibal's

perspective. Section 3 reviews other works related to

AI literacy and outlines the definition of AI selected

Curi, M. E., Capdehourat, G., Lorenzo, B., Pereiro, E. and Koleszar, V.

A Pedagogical Framework to Teach Artiﬁcial Intelligence from an Uruguayan Experience.

DOI: 10.5220/0013299900003932

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 17th International Conference on Computer Supported Education (CSEDU 2025) - Volume 2, pages 281-287

ISBN: 978-989-758-746-7; ISSN: 2184-5026

281

for this framework. Section 4 introduces the

framework itself, detailing its principles and

dimensions. Finally, the article concludes with a

reflection on the contributions of this work.

2 URUGUAYAN CONTEXT TO AI

EDUCATION FRAMEWORK

DEVELOPMENT

Launched in 2007, Uruguay's Plan Ceibal has become

a pioneering initiative by equipping every student and

teacher in the public education system - from the first

grade of primary school to the third grade of

secondary school - with technological devices like

laptops or tablets. This initiative has significantly

narrowed the national digital gap, shrinking the

technology access gap from a 13-fold difference

between the wealthiest and poorest deciles in 2007 to

just 1.2-fold by 2010, a ratio that has been sustained.

Ceibal not only distributes equipment but also

guarantees internet access across all public

educational institutions in Uruguay. It has set up a

top-notch video conferencing network across 1,650

educational centers, which covers 100% of urban

schools and benefits 97% of the public student body.

The use of advanced equipment and fiber optic

technology ensures seamless, delay-free real-time

teaching sessions, thus enhancing the delivery of

programs such as “Ceibal en inglés” (english

program) and Computational Thinking with remote

teacher participation. Uruguay's commitment to

integrating technology in education has positioned it

as the most digitized education system in the region

and has created an innovative and inclusive digital

educational landscape.

Since 2017, Ceibal and the National Education

Policy Agency (ANEP

, by its acronym in Spanish)

have been at the forefront of a groundbreaking

Computational Thinking program in Uruguay. This

initiative introduces a collaborative educational

model where a computer science-trained educator,

connected via videoconferencing, works alongside

the classroom teacher. Together, they execute a

curriculum that spans three levels, which integrates

computational thinking with a diverse range of

subjects like mathematics, science, physical

education, and language arts. This interdisciplinary

approach enriches the students' learning experience

by making the development of computational skills

relevant and dynamic. Originally, voluntary and

https://www.anep.edu.uy/

extracurricular, the program's appeal has led to

significant growth. From 30 schools at the beginning,

it has been expanded year by year. By 2024, it

reached more than 4,000 groups from 4th to 6th grade

in over 1,000 schools. This growth translates to an

educational impact on over 80,000 students and

includes more than 80% of urban public schools in

Uruguay, which highlights the program's widespread

acceptance and success.

In 2021, the curriculum evolved to incorporate

artificial intelligence, keeping pace with global

technological advancements. By 2024, to reflect its

expanded scope, the program was aptly renamed

"Computational Thinking and Artificial

Intelligence", marking a new chapter in Uruguay's

commitment to innovative and inclusive educational

practices. In the development of computational

thinking within the educational system, there is a

concerted effort to establish a comprehensive

framework for teaching and learning the theoretical

and practical aspects of artificial intelligence. The

framework presented in this article stems from the

collaboration of Ceibal’s Computational Thinking

team and the Research, Development, and Innovation

area, with the goal of establishing a framework upon

which the Uruguayan teaching community can work

within educational institutions as well as incorporate

these concepts into the classroom. This initiative goes

beyond simple tool usage and seeks to establish a

solid foundation in AI literacy. The goal is to equip

students with the knowledge and skills necessary to

face the challenges and leverage the opportunities that

AI technology will present in the near future. With an

emphasis on fostering a deep understanding of AI and

CT, the program ensures that students are prepared to

be not only proficient users but also forward-thinking

innovators in the evolving AI landscape.

3 RELATED WORK

The rapid expansion of AI in different aspects of our

lives has generated the need to prepare students to

interact effectively with this evolving technology. In

the education field, a pedagogical approach would

allow students to understand and use AI in a critical

and creative way. In this context, computational

thinking and its associated competencies have been

identified as fundamental to develop cognitive and

metacognitive skills necessary to address the

challenges of the digital age.

CSEDU 2025 - 17th International Conference on Computer Supported Education

282

3.1 Computational Thinking and

Artificial Intelligence

While the Computational Thinking Reference

Framework (Ceibal, 2022) constitutes a starting point

for the approach to AI, there are some innovations

associated with machine learning and AI that make it

necessary to extend this framework to address them

(Tedre et al., 2021). Machine learning and artificial

intelligence introduces new concepts such as data

cleaning, algorithm training and evaluation, as well as

problems associated with this process, such as bias.

These changes imply the need for a new

framework to address the issue. In this regard, Tedre

et al. (2021) propose the development of a new

computational thinking framework to approach

computational learning and AI. Although this is a

possible path, and Ceibal’s Computational Thinking

Reference Framework (Ceibal, 2022) probably

requires adjusting to the advances observed in the

subject, in this case we opted for a different

alternative. This paper proposes a more limited and

specific framework to address AI, allowing the

educational community to move forward quickly and

effectively without entering into more fundamental

reviews, which could delay the incorporation of this

subject in the short term. It is important to recognize

the deep connection between computational thinking

and AI, which materializes, for example, in the shared

principles and elements of the dimensions that make

up both frameworks.

3.2 Artificial Intelligence

Before delving into the principles and dimensions of

the framework, it is pertinent to note that the term

“Artificial Intelligence” was coined in 1956, at an

academic meeting at Dartmouth University in the

United States organized by John McCarthy and other

colleagues. As indicated in the UNESCO survey on

AI curricula (UNESCO, 2023), the definition of the

term has evolved over time, and today refers mostly

to “machines that replicate certain characteristics of

human intelligence, such as perception, learning,

reasoning, problem solving, linguistic interaction and

creative work”. It is worth noting that AI education

not only addresses learning the scientific and

technological foundations of AI, but also the

knowledge and critical reflection on how to develop

reliable AI and the consequences of not doing so

(Long & Magerko, 2020).

4 FRAMEWORK

This framework aims to become a tool to promote AI

literacy, understood as a set of competencies that

allows people to know and critically evaluate AI

technologies, use AI tools, communicate and

collaborate with AI (Long & Magerko, 2020), and

also promote different ways of thinking that

potentially allow people to create through AI.

4.1 Framework Principles

Principles are fundamental characteristics and are

incorporated transversally into all work proposals. In

this sense, they are not intrinsic elements of

computational thinking or artificial intelligence, but

guidelines for the construction of proposals and

activities to generate learning environments. The

principles recommended for implementing this

framework in the classroom are outlined below:

● Equity. Educate taking into account individual

differences and needs, without economic,

demographic, geographical, ethnic or gender

conditions impacting on students' education.

● Collaboration. Work as a team independently

and synergistically. Develop strong

interpersonal skills. Organize the group of

students to take on challenges. Make

challenging decisions and contribute to other

people’s learning.

● Creativity. Have an entrepreneurial vision, ask

appropriate questions to generate opportunities

and new ideas. Transform those ideas into

actions with a social impact.

● Autonomy. Promote exploration without fear of

making mistakes, taking risks and taking the

initiative as a strategy to actively engage in the

creation process, fostering the intrinsic

motivation of each student.

● Critical Perspective. Critically evaluate

information and arguments, identify patterns

and connections, develop meaningful

knowledge and apply it to the real world.

● Active Methodology. Use methods, techniques

and strategies that place the student at the center

of the teaching/learning process and encourage

their active participation building their own

learning experience.

4.2 Framework Dimensions

Dimensions are concepts or powerful ideas that allow

teachers to design and implement pedagogical

A Pedagogical Framework to Teach Artiﬁcial Intelligence from an Uruguayan Experience

283

proposals. They are directly connected to the skills to

AI literacy (Kim et al., 2021; Long & Magerko, 2020;

Ng et al., 2021; Olari & Romeike, 2021; Sentance &

Waite, 2002; Touretzky et al., 2019).

Figure 1: Dimensions of the AI referential framework.

We will explore each dimension, providing a

description of its scope and listing some of the

associated competencies that we consider relevant. At

the end of each dimension, we include a table (Tables

1–5) summarizing the big ideas and competencies

aligned with relevant reference literature: AI Literacy

competencies (Long & Magerko, 2020) and the five

big ideas in AI from AI4K12.org (Touretzky et al.,

2019).

4.2.1 What Is AI?

A dimension associated with the first contact with AI

is defined, which seeks to encompass the basic

concepts and the introduction to the topic. It is,

therefore, a dimension with a strong emphasis on the

identification and recognition of the subject. Taking

the AI Literacy framework as a guide (Long &

Magerko, 2020), we can identify competencies

associated with this dimension, not only by asking

“What is AI?”, but also “What can AI do?” The

approach to the subject is expected to have an

exploratory and experimental component, so it is

important to emphasize the practical nature of this

dimension, which goes beyond identifying AI. This

point is closely related to active methodologies, since

it focuses mainly on getting to know possible user

applications and experimenting with them so that the

familiarization with AI takes place through the use of

AI tools. It is also important to connect the very

definition of AI to the concepts of intelligence, the

discussion of which also enriches the debate on what

a machine must accomplish to be considered

intelligent.

The specific competencies highlighted within this

dimension are:

● Distinguish between devices that use and do

not use AI.

● To know systems that include AI components.

● Identify the properties that differentiate an AI-

based system from a rule-based system.

● Engage in learning about AI basic functions.

Understand the basics of machine learning.

● Use tools with AI. Learn about and how to use

different applications that use AI as an end

user, focusing on generative or game-related

tools.

● Critically analyze and discuss those

characteristics that make an entity

“intelligent”.

● Distinguish between general and narrow

artificial intelligence.

Table 1: The relation of the first dimension to other relevant

publications.

Big ideas AI4K12

(Touretzky et al., 2019)

Competencies AI Literacy

(Long & Magerko, 2020)

#1 Perception

#2 Representation and

reasoning

#3 Learning

#4 Natural interaction

#1 Recognize AI

#2 Understand intelligence

#4 General versus narrow

4.2.2 How Does AI Work?

There are two closely related dimensions, which we

include within the same question focused on how AI

is built. In this sense, these dimensions seek to learn

more about the fundamental concepts behind the

more technical aspects of the area, in order to use,

understand and develop AI techniques.

Representation of Knowledge. A first dimension

regarding the functioning of AI is connected to the

computational representation of knowledge. In other

words, we seek to answer the question of how

knowledge is modeled on a computer. It involves

working with data, sensors, representations and the

analysis of the human role in definitions associated

with these elements (Long & Magerko, 2020; Olari &

Romeike, 2021).

CSEDU 2025 - 17th International Conference on Computer Supported Education

284

The specific competencies highlighted within this

dimension are:

● Recognize basic concepts about data types.

● Collect relevant information from a dataset for

further processing using AI-based tools.

● Visualize data with AI algorithms.

● To know that different sensors generate

different data and to identify sensors in

different devices.

● Identify that computers perceive the world

using sensors.

● Recognize different computational

representations of knowledge and describe

some examples.

● Explain results, including errors, when

analyzing AI-provided answers, and challenge

them with questions.

● Recognize the key role that humans play in the

computational representation of knowledge in

AI-based solutions.

Table 2: The relation of the second dimension to other

relevant publications.

Big ideas AI4K12

(Touretzky et al., 2019)

Competencies AI Literacy

(Long & Magerko, 2020)

#1 Perception

#2 Representation and

reasoning

#4 Natural interaction

#7 Representations

#10 Human role in AI

#11 Data literacy

#15 Sensors

Machine Learning. The second dimension of the

framework associated with how AI works focuses on

machine learning, that is, the mechanisms that allow

a computer to learn. It involves defining a specific

task and using algorithms so that machines acquire

the knowledge necessary to carry it out successfully.

Within machine learning, data learning stands out,

which includes the analysis and processing of large

volumes of information to extract relevant patterns

and knowledge. Through machine learning,

computers are able to learn autonomously from data,

identifying regularities and generating predictive or

descriptive models. Programming plays a

fundamental role, since it allows the implementation

of learning algorithms and the development of

intelligent solutions. In addition, the human role in

the definition of tasks, the evaluation of results, the

interpretation of models and ethical responsibility in

the use of AI should be considered (Kim et al., 2021;

Long & Magerko, 2020; Ng et al., 2021; Sentance &

Waite, 2002).

The specific competencies highlighted within this

dimension are:

● Recognize that computers are able to learn

from data, including their own.

● Describe how training data can affect the

results of an AI algorithm.

● Recognize and describe examples of how a

computer reasons and makes decisions. Learn

about the simulation of the human logical

reasoning process with a computer model.

● Know the learning process of machines, as

well as the associated practices and challenges

involved.

● Recognize that computers are programmable

agents to whom it is possible to indicate the

tasks to do through a sequence of code.

● Design and program applications that use AI.

Evaluate, predict and design using AI

applications.

● Explore models created by others. Remix or

reuse code.

● Understand that humans play a key role in

programming, model selection, and fine-

tuning of AI systems.

Table 3: The relation of the third dimension to other

relevant publications.

Big ideas AI4K12

(Touretzky et al., 2019)

Competencies AI Literacy

(Long & Magerko, 2020)

#2 Representation and

reasoning

#3 Learning

#4 Natural interaction

#8 Decision making

#9 Machine learning steps

#10 Human role in AI

#12 Learning from data

#13 Critical interpretation

of data

#14 Action and reaction

#17 Programmability

4.2.3 How Should AI Be Used?

To address this question, two dimensions are

proposed. Students should understand both the

potential applications of AI and the effects and

consequences of this technology.

Computational Approach with AI. This dimension

is analogous and complementary to that of Ceibal's

Computational Thinking framework (2022) called

“computational problems”. This dimension addresses

the computational approach that involves strategies

for solving problems, such as dividing them into

parts, implementing different solutions, evaluating

A Pedagogical Framework to Teach Artiﬁcial Intelligence from an Uruguayan Experience

285

their viability and scope, in this case, particularly

involving the possibility of identifying AI strengths

and weaknesses. This involves identifying the most

appropriate problems or sub-problems to be

addressed through an AI-based solution (Kim et al.,

2021; Long & Magerko, 2020; Ng et al., 2021).

The specific competencies highlighted within this

dimension are:

● Learn about the current AI field of application:

Computer vision, speech recognition and

translation, sound, text and image generation,

among others.

● Use AI for problem solving. Apply

knowledge, concepts and applications of AI in

different scenarios.

● Recognize the type of problems where AI can

be directly applied and those that represent a

bigger challenge for AI.

● Discern when the use of AI is appropriate and

when it is best to use other tools.

Table 4: The relation of the fourth dimension to other

relevant publications.

Big ideas AI4K12

(Touretzky et al., 2019)

Competencies AI Literacy

(Long & Magerko, 2020)

#1 Perception

#2 Representation and

reasoning

#3 Learning

#4 Natural interaction

#1 Recognize AI

#3 Interdisciplinarity

#5 AI strengths and

weaknesses

#12 Learning from data

#14 Action and reaction

#17 Programmability

Ethical Use of AI and Social Impact. The ethical

and social dimension is related to the question “How

should AI be used?”, that is, to the ethical aspects and

the social impact connected to the use of AI. It is

essential to promote a critical vision and acknowledge

the impact AI has on society (Touretzky et al., 2019).

This implies a critical analysis of AI data,

understanding that it must be analyzed and interpreted

rigorously and within its context. Also,

interdisciplinary work plays a crucial role in

recognizing that there are different actors in

technology and understanding how they can

collaborate to create more complete and efficient

solutions. We also need to be able to imagine the

future of AI, exploring its potential applications and

considering its effects on the world (Kim et al., 2021;

Long & Magerko, 2020).

The specific competencies highlighted within this

dimension are:

● Identify that the use of AI has a social impact,

identifying the positive and negative effects of

AI on society and having a critical perspective

on the use of AI technology.

● Understand that data must often be analyzed

and interpreted and cannot be considered

without those processes, since AI technologies

can mirror or amplify biases, stereotypes, and

human inequalities

● Imagine the possible applications of AI in the

future and consider the effects of those

applications globally.

● Recognize collaboration with other actors,

bearing in mind that there are many different

ways of thinking and developing “intelligent”

machines.

Table 5: The relation of the fifth dimension to other relevant

publications.

Big ideas AI4K12

(Touretzky et al., 2019)

Competencies AI Literacy

(Long & Magerko, 2020)

#5 Social impact

#3 Interdisciplinarity

#6 Imagine the AI of the

future

#10 Human role in AI

#13 Critical interpretation

of data

#16 Ethics

5 FINAL DISCUSSION AND

CONCLUSIONS

According to the literature, there are few

comprehensive frameworks dedicated to teaching

artificial intelligence at the primary and secondary

levels. This paper has introduced a novel pedagogical

framework designed to promote AI literacy within the

Uruguayan educational system. The proposed

framework aims to strategically integrate AI into the

curriculum, empowering educators and education

leaders to adopt it as a foundational tool for designing

lesson plans and institutional strategies.

The framework is envisioned as a reference point

for the development of workshops, courses,

classroom activities, and educational materials. Its

flexibility and broad scope allow for adaptation

across diverse age groups and topics, making it

suitable for both primary and secondary education.

Educators can choose to address all dimensions of the

CSEDU 2025 - 17th International Conference on Computer Supported Education

286

framework or focus on specific areas depending on

the intended depth and objectives. For instance, an

introductory workshop could center only on the first

dimension, "What is AI?", while a multi-week course

could delve exclusively into the social and ethical

implications of AI, as outlined in the framework's

final dimension. These decisions involve instantiating

the framework into concrete activities by defining

specific content and pedagogical strategies.

From the Uruguayan experience, this framework

has already been utilized to design activities such as

those featured in the book "Building Artificial

Intelligence for Education" (Ceibal, 2024) and in

teacher training workshops. For example, one activity

connects the process of writing a story using a

conversational AI system to competencies promoted

within the framework’s first dimension. Other types

of experiences have been workshops of two or three

hours, both with teachers and students, where a first

approach to the subject is provided. In some cases a

first quick pass is made through all the framework

dimensions, while in other cases a specific focus has

been made on one of them.

Based on these experiences, the proposed

framework has proven to be a valuable tool for

establishing a common vision of the dimensions that

should be addressed to integrate AI into the

educational system. Feedback from both teachers and

students has been overwhelmingly positive regarding

the activities and publications developed. This

response indicates that the content is both engaging

and well-suited to the targeted educational levels.

Future efforts will focus on gathering feedback from

teaching practices to refine the framework further,

ensuring its long-term relevance and effectiveness.

In parallel with the development of this

framework, Uruguay’s educational system has

transitioned from a content-based curriculum to a

competency-based model. As part of this shift,

computational thinking has been officially

incorporated as a core competency. The existing

computational thinking framework has guided the

development of learning progressions that outline the

processes students must follow to acquire these

competencies.

The AI framework presented here serves as a

starting point to extend this work by developing

similar progressions for AI-related competencies.

Future efforts may explore the integration of the

computational thinking and AI frameworks, assessing

whether they should remain distinct or if the

computational thinking framework could be adapted

to include AI literacy competencies. These

considerations, along with the continuous refinement

of the AI framework, will be central to future research

and development in this area.

REFERENCES

Casal-Otero, L., Catala, A., Fernandez-Morante, C.,

Taboada, M., Cebreiro, B. and Barro, S. (2023). AI

literacy in K-12: A systematic literature review, in

International Journal of STEM Education, 10(1), 29.

https://doi.org/10.1186/s40594-023-00418-7 .

Ceibal (2022). Pensamiento computacional. Propuestas

para el aula. https://bibliotecapais.ceibal.edu.uy/info/

pensamiento-computacional-propuesta-para-el-aula-00

018977

Ceibal (2024). Construyendo inteligencia artificial para la

educación. https://pensamientocomputacional.ceibal.

edu.uy/wp-content/uploads/2024/06/Construyendo-Int

eligencia-Artificial-para-la-educacion.pdf

Kim, S., Jang, Y., Kim, W., Choi, S., Jung, H., Kim, S. and

Kim, H. (2021). Why and What to Teach: AI

Curriculum for Elementary School, in Proceedings of

the AAAI Conference on Artificial Intelligence, 35(17),

15569-15576.

Long, D. & Magerko, B. (April 2020). What is AI literacy?

Competencies and design considerations, in

Proceedings of the 2020 CHI Conference on Human

Factors in Computing Systems, 1-16.

Ng, D. T. K., LEUNG, J. K. L., Chu, S. K. W. and Qiao, M.

S. (2021). Conceptualizing AI literacy: An exploratory

review, in Computers and Education: Artificial

Intelligence, 2, 100041. 10.1016/j.caeai.2021.100041.

Olari, V. and Romeike, R. (October 2021). Addressing AI

and Data Literacy in Teacher Education: A Review of

Existing Educational Frameworks, in The 16th

Workshop in Primary and Secondary Computing

Education, pp. 1-2.

Sentance, S. and Waite, J. (2022). Perspectives on AI and

data science education. Recovered from

https://www.raspberrypi.org/app/uploads/2022/12/Pers

pectives-on-AI-and-data-science-education-_Sentance-

Waite_2022.pdf

Tedre, M., Denning, P. and Toivonen, T. (November 2021).

CT 2.0, in Proceedings of the 21st Koli Calling

International Conference on Computing Education

Research, pp. 1-8.

Touretzky, D., Gardner-McCune, C., Martin, F. and

Seehorn, D. (2019). AI for K-12 Envisioning: What

Should Every Child Know about AI?, in Proceedings of

the AAAI Conference on Artificial Intelligence, 33(01),

9795-9799.

UNESCO (2023). Currículos de IA para la enseñanza

preescolar, primaria y secundaria: un mapeo de los

currículos de IA aprobados por los gobiernos.

Recovered from https://unesdoc.unesco.org/ark:/48

223/pf0000380602_spa

A Pedagogical Framework to Teach Artiﬁcial Intelligence from an Uruguayan Experience

287