Future-Proofing Small Schools: Rethinking Education with AI

Giuseppina Rita Jose Mangione

1,* a

, Francesca De Santis

1,† b

Lydia Zampolini

1,‡ c

and Manuel Gentile

2,§ d

INDIRE, Firenze, Italy

ITD/CNR, Palermo, Italy

Keywords: AI Technology, Small and Rural Schools, Didactic Revitalisation, Imagination Lab, Future Proofing.

Abstract: This study investigates the transformative role of Artificial Intelligence (AI) in revitalising teaching practices

in small and rural schools, addressing their unique challenges. Combining a systematic literature review and

a participatory imagination lab (workshop) with Italian teachers, the research adopts a dual methodological

approach. The review identifies key focus areas for action, emphasising how AI can address critical issues in

“non-standard” educational contexts such as multigrade classrooms, teacher turnover, and geographical

isolation. The imagination lab complements this by exploring how these challenges are recognised in the

Italian context and what solutions are envisioned using technology cards. This participatory methodology

enables the co-design of potential AI-driven strategies tailored to real-world scenarios. The study underscores

the significance of small schools as unique laboratories for educational innovation, highlighting the

replicability and scalability of this approach. Extending such methods to a broader network of small schools

offers the potential to refine technological solutions, develop tailored intervention clusters, and foster

evidence-based, scalable policies for equitable and resilient education in similar contexts worldwide.

1 INTRODUCTION

Small and rural schools represent a worldwide

educational reality, with unique characteristics and

shared challenges that define their role within

national education systems. These institutions,

operating in a “non-standard” context, challenge

traditional school organisational models as they serve

communities in remote, mountainous, insular, or

economically disadvantaged areas, often constituting

the only educational provision available. According

to the OECD (2021), in member countries, these

schools account for approximately 20% of total

institutions, providing essential education in contexts

https://orcid.org/0000-0001-8968-3757

https://orcid.org/0009-0002-4239-643X

https://orcid.org/0009-0001-8399-2871

https://orcid.org/0000-0001-6288-0830

Author of chapters 1 and 2, and co-author of chapter 6 with

M. Gentile

†

Author of chapters 3 and 4

‡

Co-author of chapter 5 with M. Gentile

Co-author of chapter 5 with L. Zampolini and co-author

of chapter 6 with G.R.J. Mangione

characterised by low population density and limited

access to services.

The management of multigrade classes, high

teacher turnover, limited access to resources, and the

digital divide are just some of the challenges these

educational contexts face (Echazarra and Radinger,

2019).

Rural schools, as highlighted in the OECD

Learning in Rural Schools report, face significant

challenges across OECD countries. One critical issue

is resource availability: rural schools often have fewer

students, which increases per-student costs and limits

economies of scale. On average, secondary rural

schools in OECD countries have 369 students

compared to 890 in urban schools, which is

Mangione, G. R. J., De Santis, F., Zampolini, L. and Gentile, M.

Future-Prooﬁng Small Schools: Rethinking Education with AI.

DOI: 10.5220/0013410000003932

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 151-162

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

151

particularly evident in countries like Mexico,

Portugal, and the United States, where the rural-urban

difference exceeds 1,000 students. In countries like

Australia and New Zealand, up to half of primary

students in rural schools are taught in multigrade

classrooms, a necessity due to small student

populations, while in remote areas this figure rises to

90%. Similarly, in Europe, rural families often have

limited school choice; for example, in Spain, only 4%

of rural students attend private schools compared to

53% in urban areas, highlighting structural inequities.

Access to digital tools and the internet varies

significantly: in Mexico, only 42% of rural school

computers are connected to the internet, compared to

90% in urban schools. This disparity highlights the

digital divide, which exacerbates educational

inequality. Furthermore, transportation costs for rural

schools are higher, affecting access to after-school

programs and professional development

opportunities. In Italy, small and rural schools

constitute a significant part of the education system,

with over 11,600 institutions serving 48% of students

in the country’s inner areas (Bartolini et al., 2021;

2023). These schools face difficulties in ensuring

continuity of education due to teacher turnover and

the management of multigrade classes in contexts

often characterised by geographical and cultural

isolation (Mangione and Cannella, 2020), where the

risk of educational disconnection exacerbates

existing territorial disparities (Pedro et al., 2019;

UNICEF, 2020).

The COVID-19 pandemic, which highlighted the

disparity between rural and urban schools, heightened

their vulnerability. According to UNESCO (2021),

over 40% of students in global rural areas lacked

adequate resources for distance learning, leading to a

significant increase in dropout risk. In sub-Saharan

Africa, the dropout rate in rural schools rose by 15%

during the pandemic, with similar trends observed in

many parts of South Asia and Latin America (Dang

et al., 2021). Even in developed countries such as

Canada and Australia, the pandemic underlined the

urgent need for innovative strategies to ensure

educational continuity and reduce the urban-rural

divide (OECD, 2022). These numbers show how

important the issue is on a global scale and how

quickly we need to find specific ways to help small

and rural schools deal with their problems to enable

them to reach their full potential as sources of

innovation and educational resilience.

These disparities underscore the global urgency of

developing targeted interventions for rural schools,

enabling them to serve as hubs of resilience and

innovation rather than as symbols of educational

inequity. To tackle these challenges, rural education

systems must adopt innovative practices that leverage

digital tools, promote teacher training tailored to rural

needs, and strengthen local capacity. By doing so,

rural schools can transform into models of

educational equity and resilience, contributing to

broader societal progress.

Digital technologies, especially Artificial

Intelligence (AI), could represent a promising

response to these challenges. For example, could

allow small schools to overcome contextual

limitations and ensure equitable access to quality

education (Panciroli and Rivoltella, 2023). Or yet, AI

could offer significant opportunities to enhance

education by facilitating personalised learning and

inclusion. Virtual tutors and adaptive platforms could

support students in customising their learning

experience, monitoring progress in real time, and

adapting teaching strategies to specific needs

(Mangione, 2024).Thinking about the importance of

networking to overcome isolation AI could also

enable the creation of collaborative school networks,

expanding the reach of small schools through joint

projects and the exchange of educational resources or

making small schools more attractive to qualified

teachers, mitigating the problem of turnover and

ensuring greater educational continuity.

Combining cutting-edge AI technologies with

new ways of doing things could make these small

schools more competitive in the education field,

closing the gap with schools in cities and promoting

educational equality (Mangione and De Santis, 2024).

The adoption of AI would not only address contingent

challenges but could also offer an opportunity to

rethink the mission of small schools. Through the

implementation of innovative solutions, these schools

could become laboratories for educational

experimentation, promoting more equitable,

inclusive, and sustainable education (Mangione et al.

2023). In this way, AI could contribute not only to

improving teaching practices but also to

strengthening the role of small schools as drivers of

cultural and social development in their communities

(White and Corbett 2014).

2 RESEARCH QUESTIONS AND

METHODOLOGIES: BUILDING

FUTURE SCENARIOS FOR AI

IN SMALL SCHOOLS

Italian small schools, primarily located in inland,

mountainous, and insular areas, represent a “non-

CSEDU 2025 - 17th International Conference on Computer Supported Education

152

standard” educational model that addresses unique

challenges. These institutions often serve as the only

educational hubs within local communities, playing a

crucial role in maintaining social and cultural

cohesion in these territories. According to Bartolini et

al, 2023, Atlas of Small Schools, Italy has over 11,600

small schools, catering to approximately 48% of all

students in the national educational system,

demonstrating the pervasive and persistent relevance

of this phenomenon (Bartolini et al., 2021).

For years, INDIRE has been contributing to the

improvement of educational offerings and teaching-

learning experiences in non-standard educational

situations through continuous service research

(Cannella, Mangione and Rivoltella, 2021;

Mangione, 2024a). This approach views research as a

service to educational processes, addressing

criticisms of educational research institutions for their

alleged “failure to respond to the demands of schools”

or “institutional disengagement.”

By doing research in these contexts, Italian small

schools become an epistemic context—a privileged

space to identify research questions and test the

outcomes of investigations, contributing to the

advancement of pedagogical science and enhancing

the most fragile educational realities (Mangione,

2024b).

2.1 The Research Questions

We identified two complementary research questions

within this framework.

The first focuses on the analysis of the existing

“scientific discourse” concerning AI in rural

educational contexts: RQ1. What are the primary

application domains of AI in rural educational

contexts?

This question seeks to explore the dimensions

identified in the literature as fundamental for applying

AI in remote, non-standard contexts characterised by

limited access to resources. The necessity of

revitalising educational practices in these contexts is

a globally recognised issue, as highlighted in

international reports (UNESCO, 2021; Trendov,

Varas and Zeng, 2019), which emphasise the role of

AI in overcoming territorial disparities and

introducing innovative teaching practices in

geographically and culturally isolated schools.

The second question shifts from the theoretical

dimension to the local one, contextualising the global

reflections into the specific context of Italian small

schools: RQ2. What are the problem scenarios

specific to Italian small schools, and what AI-driven

solutions can be proposed to address them?

This question aims to relate global evidence to

situated problems. The goal is to understand how AI

technologies can be designed or adapted to address

concrete challenges, such as managing multigrade

classes, ensuring educational continuity amid teacher

turnover, and mitigating the risk of social exclusion

for students in marginal areas (Mangione and

Cannella, 2020; Mangione, 2023). This step is crucial

to orient research towards identifying technologies

that address existing problems and radically rethink

educational experiences, imagining future scenarios

and innovative solutions.

The duality of these questions is not only

methodological but also epistemological: on one hand,

the aim is to understand whether and how AI can

already be considered a strategic opportunity for rural

and marginal educational contexts; on the other hand,

the goal is to ground this reflection in a participatory

process involving teachers and local stakeholders to

define realistic and sustainable use cases.

2.2 Research Methodology

The research questions require a methodology that

alternates theoretical analysis with practical

experimentation to root technological innovation in

concrete contexts, avoid standardised approaches,

and promote solutions that respect the peculiarities of

small schools.

In the first phase - to address the first research

question (RQ1) What are the primary application

domains of AI in rural educational contexts? - a

scoping review was conducted. This methodology is

particularly suited to providing a comprehensive

overview of a broad topic such as AI in rural education

(Peterson et al., 2017). Following the model proposed

by Arksey and O’Malley (2005) and previously

applied by Mangione and De Santis (2024), the

scoping review was developed through five phases.

After identifying the research question, “What does

the literature say about AI and rural education?”, a

secondary question was defined to identify the main

application domains of AI in rural education contexts.

We identified studies using databases such as Web of

Science, Scopus, and Google Scholar. Initial

keywords included “artificial intelligence” and “rural

education”. Subsequently, we expanded the query to

include terms such as “machine learning”, “deep

learning”, “artificial education”, “rural school”, and

“small school”. Inclusion criteria required that studies

be published in English, be open-access, and be dated

from 2010 onward. The scoping review, which

included a quantitative analysis and a thematic

summary of AI’s main uses, led to the focus being put

Future-Prooﬁng Small Schools: Rethinking Education with AI

153

on how AI can be used to improve teaching in rural

schools (Mangione and De Santis, 2024). After that,

the investigation proceeded with what Arksey and

O’Malley (2005) identify as the sixth and optional

final phase of a scoping review: consulting

stakeholders to offer additional sources of

information, perspectives, meanings, and

applicability. Through a “spoken reflection” process

involving national and international experts selected

for their knowledge of small school contexts,

opportunities for small schools were identified by

connecting the stimulus questions to the dimensions

emerging from the scoping review, converging on

specific challenges for revitalising teaching in small

schools.

In the second phase - to address the second

research question (RQ2) What are the problem

scenarios specific to Italian small schools, and what

AI-driven solutions can be proposed to address them?

- a participatory workshop was conducted based on

the Design Thinking (DT) methodology (Brown,

2009). This approach, increasingly used in

educational research, is particularly effective for

addressing complex problems and developing

innovative solutions in contexts characterised by

structural constraints, such as small schools (Razzouk

and Shute, 2012). A key aspect of the workshop was

the use of technology cards to foster imaginative,

project-based thinking and define future scenarios

(Sanders and Stappers, 2014).

Engaging 46 teachers

from primary and lower secondary schools, the

activities followed the DT

stages. The participants

began by sharing their experiences and challenges

through a story-share and capture tool, fostering

empathy and highlighting issues like multigrade class

management, geographic isolation, and teacher

turnover. Building on this, they collaboratively

defined problems using “How Might We…?”

(HMW) questions to frame challenges constructively,

setting the stage for the ideation phase. The

introduction of technology cards, describing current

or potential AI technologies and their educational

applications, encouraged innovative thinking. They

have been defined in the context of the AI-wareness

board game to engage teachers in initial thinking

about the use of AI in school settings (Re et al., 2024).

This approach, recognised in participatory design

literature (van Amstel et al., 2012), enabled the

exploration of creative solutions tailored to the

challenges identified, proving particularly useful in

educational research (Wölfel & Merritt,

2013).Through this process, the workshop not only

identified potential applications of AI but also laid the

groundwork for a transformative vision that

reimagines the role of emerging technologies in

addressing the unique challenges of small schools.

The results of these methodologies go beyond

identifying AI applications. They propose a

transformative vision capable of fundamentally

rethinking the role of emerging technologies in the

context of small schools. The results will be presented

in the subsequent sections.

3 AI IN RURAL SCHOOLS:

INSIGHT FROM THE SCOPING

REVIEW

The scoping review led to the realisation of an

exploratory study, which focused on the 19 studies

that met the inclusion criteria. After analysing the

studies, we were able to map them by examining

recurring themes and subthemes.

3.1 Key Themes and Areas of Focus

The review identified four main thematic clusters

(Table 1).

Most of the papers (12) fall into cluster 1, “AI for

revitalising teaching and learning processes”. These

studies focus on exploring the applications of AI in

the field of education, with the main objective of

reducing the existing disparities in the quality of

education between urban and rural contexts. In

particular, the areas of application include the use of

AI to personalise learning, improve disciplinary

teaching and distance learning processes, and to

integrate learning AI in K-9 and K-12 syllabi to create

engaging learning experiences.

Other studies (3) can be attributed to cluster 2, “AI

for teacher professional development”. This research

is based on the premise that it is necessary to

overcome the professional isolation of teachers, which

often characterises rural contexts, focusing on teacher

training as a lever to improve access to educational

resources and to promote equity (Mangione, Pieri and

De Santis, 2023). Some studies explore AI’s potential

to enhance resource sharing and peer collaboration

(Wang, 2020), while others focus on teachers’

perceptions and ethical challenges (Chounta et al.,

2022). Edwards and Cheok (2018), propose AI and

robotics as tools to address teacher shortages, ensuring

teachers are prepared to integrate these technologies

while maintaining their role in social interaction and

emotional support (Gentile et al., 2023).

The contributions (2) attributable to cluster 3, “AI

for developing predictive models of student interest

CSEDU 2025 - 17th International Conference on Computer Supported Education

154

and success”, highlight the potential of AI to

construct predictive models that enhance educational

outcomes by personalising learning and providing

targeted student support.

The studies (2) categorised under cluster 4, “AI

for school service management and risk prediction”,

focus on the application of AI to enhance

administration, optimise transportation, and improve

safety and accessibility in disadvantaged or isolated

schools.

It is important to highlight that the identified

themes and sub-themes are interconnected, forming a

complex yet integrated ecosystem for potential

applications of AI in education. For instance, the

implementation of AI-based personalised learning

tools (a sub-theme of the first cluster) requires

adequate training for teachers (the second cluster).

Teachers must acquire specific skills to utilise these

technologies and adapt them to their educational

contexts. Moreover, these tools generate a significant

volume of student data, enabling the creation of

sophisticated predictive models (the third cluster). A

fragmented approach would risk reducing the positive

impact of technology and widening existing

inequalities, especially in rural or resource-limited

contexts. Therefore, an integrated strategy that fosters

collaboration between researchers and teachers is

essential to ensure that solutions are pedagogically

relevant and socially inclusive.

3.2 The Studies Guiding the Design of

the Participatory Workshop

The studies most closely related to the teaching

context, especially those from cluster 1 and some

from cluster 2, guided the research group in the next

phases and the design of the participatory workshop

for small-school teachers. This analysis provided

insight on how to effectively use AI to address the

unique needs of this educational setting. Many

sources highlight the AI’s potential to personalise

learning experiences. For example, Yang and Zheng

(2021) argue that, despite existing economic

challenges, AI presents a concrete solution to reduce

inequalities in the distribution of educational

resources. This technology can offer students in

remote areas of China the chance to expand their

knowledge and horizons. One advantage of AI, as

highlighted by the authors, is the possibility of

providing personalised instruction by analysing

students’ progress data. This analysis helps identify

the so-called “dead zones” in their learning and

allows educators to respond in a targeted way to

individual needs. Wang and Lin (2019) also underline

how AI and big data are transforming education by

fostering personalised, ubiquitous, and lifelong

learning.

Table 1: Reference distribution: themes and sub-themes

Themes Sub-themes References

AI for revitalising

teaching and learning

processes

- AI for the personalisation of learning

- Intelligent Tutoring Systems

- Automated assessment tools

- Integrating AI into K-9 and K-12 education

- AI to improve distance and disciplinary teaching

Gong et al. (2023)

Gong et al. (2020)

Iyer (2022)

Iyer (2019)

Jiang (2021)

Jiang et Cheong (2023)

Rasheed et al. (2021)

Vanderberg et al. (2022)

Wang and Lin (2019)

Xiao et al. (2022)

Yang and Zheng (2021)

AI for teacher professional

development

- Teachers’ knowledge, perceptions and attitudes

towards AI and ethical challenges

- AI for the continuous professional

development of teachers

- Teacher trainin

on the use of AI

Chounta et al. (2022)

Edwards and Cheok (2018)

Wang (2020)

AI for developing

predictive models of

student interest and

success

- Predicting student interest in higher education

(orientation)

- Early identification of learning difficulties and

personalised support

Nuankaew and Nuankaew (2022)

Saravanan et al. (2021)

AI for school service

management and risk

prediction

- Transport optimisation

- Assessment of the vulnerability of schools to

natural events

- O

timisation of resource allocation

De Souza Lima et al. (2023)

Yousefi et al. (2020)

Zhou (2022)

Future-Prooﬁng Small Schools: Rethinking Education with AI

155

The implementation of personalised education,

which is not possible in traditional education due to

limitations like teacher shortages, can help address the

imbalance in educational resources and help address

social problems by creating a new ecology of

educational technology. Chounta et al. (2022) present

a study on Estonian teachers regarding the use of AI

as a tool to support teaching in K-12 schools. The

research is interesting because it reveals that teachers

have a positive view of AI’s educational potential,

particularly its ability to personalise learning. In their

paper, the authors cite Intelligent Tutoring Systems

(ITS) as a relevant example of how AI can improve

access to quality education, especially in rural areas.

ITS are characterised by their capacity to create

personalised learning paths that adapt to each

student’s pace and learning style. In addition, ITS are

valuable tools for accessing, adapting and using

multilingual content. Their advanced natural language

processing capabilities can enhance learning in

multilingual environments, enabling students to

interact with the system in their native languages. This

approach makes the educational experience more

inclusive and accessible. For example, AI could be

integrated into a Learning Management System

(LMS) to automatically translate content, allowing

students from diverse language backgrounds to easily

access information. Iyer (2019) also analyses the

Indian context to highlight the complex challenge

represented by linguistic diversity in education. With

many regional languages, India faces significant

challenges in creating and distributing educational

content that is accessible to all students. In response to

this issue, the research suggests leveraging AI as an

innovative solution by integrating automatic

translation systems into mobile devices for learning.

These systems would allow students to access

translations of words or sentences anytime, making it

easier to understand teaching materials and promoting

inclusion. Other researchers (Edwards and Cheok,

2018) propose the use of robots as teachers, especially

in situations where there is a shortage of teaching staff.

In their paper, the authors present a project aimed at

developing a prototype teacher robot and outline its

potential capabilities for delivering educational

content and facilitating social interactions. The article

also discusses the ethical and technological challenges

associated with this concept, considering the

conflicting opinions on whether robots can fully

replace human teachers; the authors suggest future

research directions to address AI’s current limitations

in the educational field. Another theme emerging from

the studies is the use of AI to develop more efficient

assessment systems that address students’ different

backgrounds and learning styles. AI can automate

repetitive assessment tasks, such as marking multiple-

choice quizzes or analysing short answers, which

saves teachers valuable time for more complex

responsibilities, such as providing individual support

to students (Yang and Zheng, 2021). Furthermore, AI

can offer students immediate feedback during

exercises or activities, helping them to identify and

correct errors in real-time (Chounta et al., 2022). This

timely feedback fosters a deeper understanding of

concepts and more active learning. The analysed

contributions propose a critical reflection regarding

the integration of AI in evaluations, which should be

accompanied by strategies aimed at preventing

systemic biases and promoting a collaborative

approach in which AI supports, but does not replace,

human experience. In other studies (Gong et al., 2020;

Vandenberg et al., 2022), research on AI-based

education focuses on its integration into national

curricula. For example, Gong et al. (2020) conducted

an in-depth study on the current state of AI education

in Qingdao, China. One of the main goals is to develop

a curriculum that integrates the concepts and

applications of AI across disciplines, providing

students with a comprehensive understanding of AI

and its impact on society. This need is particularly

relevant in rural areas, where educational resources

are often limited, and opportunities for AI learning are

fewer compared to urban settings. Vandenberg’s

(2022) presents research that investigates the potential

of video games to enhance interest and promote

understanding of AI and computing concepts.

Through game design activities, students can learn the

basic principles of AI in a practical and creative way,

promoting active engagement and meaningful

learning. The analysis of the contributions that

emerged from the scoping review highlighted how the

solutions and these technological tools appear to be

promising resources for addressing the specific and

complex challenges that characterise these

environments and offer innovative opportunities for

improving the effectiveness and equity of education in

these contexts.

4 THE EXPERTS’

PERSPECTIVE: CHALLENGES

AND OPPORTUNITIES FOR

SMALL SCHOOLS

The themes from the scoping review guided the

development of questions for participatory interviews

with Italian and international experts about small

CSEDU 2025 - 17th International Conference on Computer Supported Education

156

schools. The interviews revealed a consensus on the

importance of AI as a tool to reduce the educational

gap between urban and rural schools (Mangione and

De Santis, 2024). Experts identify that AI provides

real-time feedback to teachers, allowing immediate

changes in their teaching methods. AI can also be

combined with gamification, which makes learning

more engaging and tailored for students, helping them

stay motivated and included. Another key point was

AI’s role in breaking down language barriers by using

translation technology. This makes educational

materials accessible to students who speak different

languages and encourages schools in different

countries to work together. AI also supports teachers

in managing classrooms with students of different

ages and levels (Perna, et al. 2024). If a teacher is

absent, AI can provide online resources, so education

continues. Additionally, AI can strengthen

connections in isolated school communities by

creating networks that involve everyone. Overall,

experts agreed that AI should be a helpful tool in

education, not a replacement for teachers (Gentile et

al., 2023). Its positive impact depends on focusing on

students’ needs and recognising the important role

teachers play in education.

5 IMAGINING FUTURE

SCENARIOS: THE WORKSHOP

The workshop “Future Scenarios: AI for Small

Schools” engaged 46 teachers to explore how AI can

personalise learning and enhance teaching, focusing

on bridging urban-rural gaps. The participants were

divided into six groups based on school grades, to

address common challenges and opportunities. A DT

approach was adopted, combining theory and

collaboration to co-construct future scenarios

addressing AI in “non-standard” education, focusing

on the empathy, definition, and ideation phases, thus

encouraging critical reflection on AI’s benefits and

challenges (d.school, 2010; d.school, 2018).

5.1 Phase 1: Empathy

The empathy phase explored challenges in small

schools by leveraging teachers’ expertise to identify

common issues and potential solutions. To facilitate

the process, participants were given an adapted

version of the story-share and capture tool (d.school.

2010), focusing on context, actors, needs, and

problems. This collaborative approach facilitated

narration, active listening, note-taking, and the

synthesis of shared insights. During this phase, each

group member shared their observations, highlighting

quotes, surprises, and significant details through post-

it notes, which were then grouped to identify common

themes and patterns. The ultimate goal of this activity

was to fully understand their experiences, find out

their needs in relation to the research theme, and thus

initiate a collective reflection on possible solutions.

5.2 Phase 2: Definition

Each group of participants analysed and discussed the

main challenges that emerged from their shared

stories (Table 2). This step favoured a common

understanding of the context and priorities,

highlighting the similarities between the experiences

that arose. Starting from the themes which emerged,

common elements were identified between the

different schools, leading to the creation of a shared

scenario helping to narrow down the problem.

This process enabled the framing of the ‘How

Might We...?’ (HMW) question, a fundamental step

that translated the identified challenges into a

constructive and creative question. The HMW

question is a short question that launches the ideation

phase and is formulated in such a way that it is broad

enough to include a wide range of solutions, but at the

same time focused enough to provide useful

boundaries (d.school, 2018). The question, aimed at

circumscribing the problem in a constructive and

creative way, helps participants in the process of

translating challenges into questions that open up

problem-solving, leading to the next stage of the

workshop. Group 1 worked on schools located in areas

with poor connections, characterised by multi-grade

or overcrowded classes. In this context, low

motivation was identified as the main problem. This

phenomenon is associated with geographical

isolation, which limits access to diversified

educational experiences, and the complexity of

managing heterogeneous classes. Emerging needs

include training teachers and offering students

customised orientation paths to stimulate their interest

and active participation. The HMW question

formulated by the group is: “HMW increase

motivation?”. This question highlights the importance

of simultaneously addressing motivational,

pedagogical, and organisational aspects.

Group 2 focused on secondary school and the need

for interventions in students and teachers’ conscious

and correct use of AI. The problem limits the

educational potential of AI and creates inequalities

in access to educational opportunities. The group

highlighted

the need to develop training paths for all

Future-Prooﬁng Small Schools: Rethinking Education with AI

157

Figure 1: Poster layout for the empathy and definition

phase.

school stakeholders (students, teachers, and families)

and to improve school infrastructure to support a more

effective and inclusive use of technology. The HMW

question summarising these needs is: ‘HMW develop

an informed use of AI to support learning?’. This

perspective emphasises the need to combine

technological innovation with cultural education.

Group 3 analysed the context of an

omnicomprehensive institute with a multi-grade class

from first to fifth primary, in which more than 30% of

the students are of foreign origin. The main problem

concerns learning Italian as a second language (L2),

a barrier that hinders both school success and social

inclusion among foreign students. The group

identified the need to increase the number of teachers

supporting the classes and to introduce technological

tools to facilitate language learning. The HMW

question is: ‘HMW use language mediators supported

by AI?’. The answer to this question requires an

integrated approach that combines technological and

pedagogical solutions to foster language inclusion.

Group 4 focused its reflection on challenges

related to educational poverty, analysing a school

context characterised by frequent teacher turnover,

educational discontinuity and a strong need for

affective and social inclusion. It emerged that

meeting these needs, requires stable and inclusive

educational pathways supported by technological

tools. The HMW question posed by the group is:

‘HMW ensure continuity?’. This question invites

reflection on the need for a structural intervention to

promote stability and inclusion.

5.3 Phase 3: Ideation

Following the HMW question, the ideation phase

encouraged broad exploration of solutions through

brainstorming, emphasising creativity and

collaboration (d.school, 2010; d.school, 2018). Using

the ‘Yes, And!’ method, participants built on each

other’s ideas to foster innovation. Technology cards

from the AI-wareness board game (Re et al., 2024)

served as creative constraints, inspiring AI-driven

solutions tailored to educational challenges. With this

approach, the groups began to develop ‘game-

changing’ ideas that could lead to transformations by

introducing AI. The solutions that emerged from this

co-creation process were not limited to individual

proposals; in fact, each group could select 3

technology cards that would lead to the construction

of a future scenario (Table 3). This scenario provided

a clear and inspiring vision, outlining how the school

environment could evolve through the strategic use of

AI.

Table 2: Group results.

Group Context Problems Needs HMW

1 Isolation

Multi-grade classes

Low student motivation Teacher

training

Personalised

guidance

HMW increase motivation?

2 Secondary school

Conscious use of

technology

Unconscious and incorrect use

of AI

Training for

an informed

use of AI

HMW develop a conscious use

of AI to support learning?

3 Multi-grade class

Foreign students

Italian L2 Increased

number of

teachers

HMW use linguistic mediators

supported by AI?

4 Area with educational

poverty

Lack of educational continuity

Frequent teacher turnover

Prejudices

Social and

cultural

inclusion

HMW ensure continuity?

CSEDU 2025 - 17th International Conference on Computer Supported Education

158

Figure 2: Poster layout for the ideation phase.

Table 3: Technology cards chosen by each group.

Group Card 1 Card 2 Card 3

1 Intelligent

LMSs

Multimodal

Intelligent

Assistants

Intelligent

Tutoring

Systems

2 Intelligent

Tutoring

Systems;

LLM-based

chatbots

Intelligent

classrooms;

other

technology

Expert

Systems

3 Multimodal

Intelligent

Assistants

Intelligent

Tutoring

Systems

4 Intelligent

Tutoring

Systems

Writing

assistants

AI-based

maths

solvers

Group 1 focused on increasing motivation in

isolated school contexts with heterogeneous classes,

linking motivation problems with the potential

offered by the chosen technologies. Starting with

intelligent LSMs, intelligent multimodal assistants,

and intelligent tutoring systems, the imagined future

scenario proposes an integrated learning support

system designed to increase motivation. Classes,

although heterogeneous, benefit from intelligent tools

that personalise the educational experience and offer

innovative solutions to logistical and motivational

challenges. Thanks to these technologies, multi-grade

schools turn into innovative laboratories for

experimenting with new educational methodologies.

The school community of motivated students and

teachers is an example of how AI can turn educational

challenges into opportunities, reducing inequalities

caused by geographical isolation. It is an approach

that aims to increase student motivation (with more

engaging and relevant learning experiences), support

teachers in managing heterogeneous classrooms, and

foster inclusion by overcoming geographical and

logistical barriers through technology.

Group 2 focused on the use of AI to improve

administration and promote technology-aware use in

schools. The chosen cards included intelligent

tutoring systems, LLM-based chatbots, intelligent

classrooms, and expert systems. The group imagined

a future in which secondary school, starting from a

situation of infrastructural and cultural difficulties

concerning the use of AI, becomes a role model for

digital awareness. By integrating AI to support

administrative and teaching activities, a systemic

change that involves students, teachers, and families

is promoted. The group proposed the use of AI tools

to improve the search for funding and calls for

tenders, making it easier to find economic resources,

often not intended for small schools. The question

‘HMW develop an informed use of learning support?’

highlights a seeming discrepancy with the proposed

future scenario in which AI is used to facilitate the

search and finding of tenders and funds in

institutional portals to support planning at the

administrative level. However, the two plans find a

meeting point insofar as finding calls and funding

could have an immediate impact by combining two

fronts: on the one hand, AI tools to improve

administrative management, and on the other,

training and workshops to create a digitally aware

culture. Therefore, it serves as an illustration of how

a systemic and collaborative vision can effectively

tackle a complex problem.

Group 3 identified an emerging need for increased

support in the classroom, proposing the adoption of

language mediators supported by AI as a solution.

The future scenario foresees the use of multimodal

assistants to create interactive learning experiences,

such as simulations and multimedia activities, that

help students improve language skills. Intelligent

tutoring systems would offer customised paths based

on the specific needs of each pupil, promoting not

only language learning but also integration into the

school environment. This system improves students’

language skills and fosters greater social cohesion

leading to a more inclusive school environment. The

future scenario imagines a school in which AI does

not replace the role of teachers but supports them in

responding to complex challenges.

Group 4 emphasised the need to ensure

educational continuity in a school characterised by

educational poverty and a frequent teacher turnover.

The selected technologies include intelligent tutoring

systems, writing assistants, and AI-based

mathematics solvers. The imagined scenario involves

an AI-based tutoring system, which by creating an

‘educational memory’ ensures continuity in learning

paths, even when teachers change. Writing assistants

Future-Prooﬁng Small Schools: Rethinking Education with AI

159

and maths solvers have been proposed as practical

tools to improve students' academic skills, with a

focus on inclusion, support and educational stability.

In the future scenario, with these solutions, the school

becomes a model for educational innovation in

contexts of social and cultural poverty. The intelligent

tutoring system improves educational continuity and

creates an inclusive school environment where

students and teachers feel supported and motivated.

The collaboration between schools, families, and the

territory is strengthened, creating a cohesive and

resilient educational community.

6 DISCUSSION AND RESEARCH

PERSPECTIVES

The research presented represents a pioneering

contribution to revitalising teaching in small and rural

schools, proposing an approach that combines

scientific rigour with active teacher participation.

Small schools, as highlighted in the international

literature, represent a privileged context for

experimenting with educational innovations due to

their ability to adapt to territorial specificities and

local challenges (Bartolini et al., 2021; Mangione et

al., 2023). However, these institutions often face

critical challenges, such as managing heterogeneous

and multigrade classes, teacher turnover, and

geographical isolation (OECD, 2020; Corbett &

White, 2014). Within this context, the revitalisation

of teaching, understood as the adoption of innovative

pedagogical approaches supported by AI, emerges as

a crucial strategy.

At the same time, while the introduction of AI in

small schools offers promising opportunities, it is

essential to acknowledge the challenges related to its

implementation. Limited access to digital

infrastructure, unstable broadband connectivity, and

the need for targeted teacher training represent

significant barriers. AI should not be viewed as a

universal solution to educational challenges but rather

as a tool that, when properly contextualized, can

complement and support existing pedagogical

practices.

The systematic review of the literature conducted

in this research identified four main clusters of AI

applications in educational contexts, with the cluster

dedicated to the revitalisation of teaching offering the

most relevant evidence for the Italian pilot case.

Recent studies demonstrate how AI-based tools, such

as ITS and adaptive platforms, can improve access to

educational resources and effectively personalise

learning, particularly in resource-constrained

contexts (Yang and Zheng, 2021; Xiao et al., 2022).

The DT methodology used in the workshops

effectively addressed the complexity of small

schools’ educational challenges. Structured into

empathy, definition, and ideation phases, the process

identified priority problems and explored realistic

applications of AI technologies in education. The use

of technology cards was particularly significant,

facilitating design imagination by providing concrete

insights into existing or hypothetical technologies and

promoting a critical and informed reflection on AI’s

potential (van Amstel et al., 2012; Wölfel and Merritt,

2013). The technologies selected by the pilot groups,

such as intelligent LMSs, adaptive virtual tutors, and

multimodal assistants, not only address the needs

identified by teachers but also align closely with

evidence from international literature. For example,

ITS has been recognised as an effective tool for

personalising learning and supporting teachers in

managing heterogeneous classes (Chounta et al.,

2022). Similarly, the use of AI-based linguistic

mediators, proposed to address language learning

barriers, aligns with studies highlighting how

translation and language recognition technologies can

promote social inclusion and reduce educational

inequalities (Iyer, 2019; Jiang, 2021). Integrating

these technologies into school contexts represents an

educational transformation strategy that goes beyond

solving contingent problems, laying the foundation

for a new vision of schools as laboratories of

innovation and community resilience.

From a methodological perspective, a crucial

aspect for future research is the need for scaling up

the experimental approach adopted. The imagination

lab conducted in the Italian pilot context offers a

replicable model that could be extended to other small

schools within the national network, creating an

ecosystem of widespread experimentation. This

process would not only validate the technologies and

scenarios imagined but also identify specific

technological clusters to address recurring problems.

For example, further experimentation could explore

the combined potential of ITS and multimodal

assistants in multilingual contexts or the role of

intelligent LMSs in ensuring educational continuity

in areas affected by high teacher turnover.

Additionally, we aim to scale up the lab and extend it

to multiple pilot groups within the national network

of small schools, leveraging a constantly updated

clustering approach provided by the Atlas of Small

Schools by Bartolini et al. (2023). Expanding the

lab’s reach could also generate comparative data that

would support the development of a pedagogical

CSEDU 2025 - 17th International Conference on Computer Supported Education

160

framework integrating DT, technological

experimentation and active participation. This

initiative would pave the way for scalable educational

policies grounded on robust scientific evidence.

Extending this lab nationally and internationally

could enable the development of a map of emerging

technologies capable of responding to diverse

educational scenarios. Future scientific research

should focus on analysing the dynamics of

implementing these technologies organised by

problem scenarios in complex educational contexts,

monitoring results, and proposing improvements

based on empirical data. Building regional

technology clusters, as suggested by Mangione et al.

(2023), would be a crucial step in transitioning from

experimentation to the realisation of scalable

educational models that address local challenges

while generating transferable knowledge to influence

educational systems on a larger scale.

REFERENCES

Arksey, H., & O’Malley, L. (2005). Scoping studies:

Towards a methodological framework. International

Journal of Social Research Methodology, 8(1), 19–32.

https://doi.org/10.1080/1364557032000119616

Bartolini, R., Mangione, G.R.J., De Santis, F., & Tancredi,

A. (2021). Piccole scuole e territorio: Un’indagine sulla

relazione scuola-comune per un progetto formativo

allargato. Scienze del Territorio, 9, 155–167.

https://doi.org/10.13128/sdt-12259

Bartolini, R., Mangione, G.R.J., & Zanoccoli, C. (2022).

Small schools: Rethinking the forme scolaire for an

educational compact that extends to the community and

the territory. Formazione & Insegnamento, 20(2), 14-

35.

Bartolini, R., Zanoccoli, C., Mangione, G.R.J. (2023).

Atlante delle piccole scuole in Italia. Mappatura e

analisi dei territori con dati aggiornati all’anno

scolastico 2020/21. INDIRE, Firenze.

Brown, T. (2009). Change by design: How design thinking

creates new alternatives for business and society.

Harper Business.

Cannella, G., Mangione, G.R.J., & Rivoltella, P.C. (2021).

A scuola nelle piccole scuole. Storia, metodi,

didattiche. Brescia: Morcelliana Scholè.

Chounta, I.A., Bardone, E., Raudsep, A., & Pedaste, M.

(2022). Exploring teachers’ perceptions of artificial

intelligence as a tool to support their practice in

Estonian K-12 education. International Journal of

Artificial Intelligence in Education, 32(3), 725–755.

White, S., & Corbett, M. (2014). Doing educational

research in rural settings. Methodological issues,

international perspectives and practical solutions.

Routledge.

De Souza Lima, F. M., Pereira, S. D., Da Conceição, S. V.,

& De Camargo, R. S. (2017). A multi-objective

capacitated rural school bus routing problem with

heterogeneous fleet and mixed loads. 4OR-Q Journal of

Operational Research, 15(4), 359–386.

d.school (2010). An introduction to design thinking process

guide. Hasso Plattner Institute of Design at Stanford.

Retrieved from https://dschool.stanford.edu

Echazarra, A. and T. Radinger (2019). Learning in rural

schools: Insights from PISA, TALIS and the literature,

OECD Education Working Papers, No. 196, OECD

Publishing. https://doi.org/10.1787/8b1a5cb9-en.

Edwards, B. I., & Cheok, A. D. (2018). Why not robot

teachers: Artificial intelligence for addressing teacher

shortage. Applied Artificial Intelligence, 32(4), 345–

360.

Gentile, M., Città, G., Perna, S., & Allegra, M. (2023). Do

we still need teachers? Navigating the paradigm shift of

the teacher’s role in the AI era. In Frontiers in

Education (Vol. 8). Frontiers Media SA.

https://doi.org/10.3389/feduc.2023.1161777

Gong, X., et al. (2023). Board 433: Work in progress:

Building a “Project-based learning for rural Alabama

STEM middle school teachers in machine learning and

robotics” RET site.

Gong, X., et al. (2020). K-9 artificial intelligence education

in Qingdao: Issues, challenges and suggestions. In 2020

IEEE International Conference on Networking, Sensing

and Control (ICNSC), Nanjing, China, pp. 1–6.

Iyer, S. (2022). Robust model for rural education using deep

learning and robotics. In Troiano, L., et al. (Eds.),

Advances in Deep Learning, Artificial Intelligence and

Robotics. Lecture Notes in Networks and Systems, vol

249. Springer, Cham.

Iyer, S. (2019). Futuristic technology and innovation to

solve the problems of Indian rural education.

International Journal for Research in Applied Science

and Engineering Technology, 7(12), 920–930.

Jiang, C. (2021). Technical framework and model of

artificial intelligence for boosting the revitalization of

rural education. Proceedings of the 2nd International

Seminar on Artificial Intelligence, Networking and

Information Technology (AINIT), pp. 107–110.

Lor, R. (2017). Design thinking in education: A critical

review of literature. Asian Conference on Education &

Psychology, Bangkok, Thailand, 24–26 May.

Mangione, G.R.J. (2024a). The small and rural school

Italian network for digital technologies flourishing in

“non standard” educational context: Before, during, and

after the pandemic era. Disruptive Technologies in

Education and Workforce Development (pp. 136–164).

IGI Global.

Mangione G.R.J. (2024b) Rete per le piccole scuole. Tra

cultura dell’educazione e Service Research, In G.

Pastori, L. Zecca, & F. Zuccoli (eds), Cantieri aperti e

scuole in costruzione: Alla ricerca di nuovi “modelli”

e pratiche per una scuola democratica. Milano:

FrancoAngeli.

Mangione, G. R. J., & Cannella, G. (2021). Small school,

smart schools: Distance education in remoteness

Future-Prooﬁng Small Schools: Rethinking Education with AI

161

conditions. Technology, Knowledge and Learning,

26(4), 845-865.

Mangione, G. R. J., & De Santis, F. (2024). Artificial

intelligence in non-standard educational contexts:

Results from a “spoken reflection” on small schools.

Scholé: Intelligenza Artificiale ed Educazione, 1/2024.

Nuankaew, P., & Nuankaew, W. S. (2022). Student

performance prediction model for predicting academic

achievement of high school students. European Journal

of Educational Research, 11(2), 949–963.

Panciroli, C., & Rivoltella, P. C. (2023). Pedagogia

algoritmica. Per una riflessione educativa

sull’Intelligenza Artificiale (pp. 1-240). Scholé-

Morcelliana.

Pedro, F., Subosa, M., Rivas, A., & Valverde, P. (2019).

Artificial intelligence in education: Challenges and

opportunities for sustainable development. UNESCO

Publishing.

Perna, S., Allegra, M., Gentile, M., Ottaviano, S., Pitrella,

V., Re, A., Tosto, C., & Città, G. (2024). Analysing the

impact of Artificial Intelligence on teacher-student

interaction in the light of actor-network theory. Scholé:

Rivista Di Educazione e Studi Culturali, LXII, 1, 125-

145.

Rasheed, M. A., et al. (2021). Use of artificial intelligence

on electroencephalogram (EEG) waveforms to predict

failure in early school grades in children from a rural

cohort in Pakistan. PLOS ONE, 16(2).

Razzouk, R., & Shute, V. J. (2012). What is design thinking

and why is it important? Review of Educational

Research, 82(3), 330–348.

https://doi.org/10.3102/0034654312457429

Re, A., Città, G., Gentile, M., Presti, F. L., Pitrella, V.,

Tosto, C., & Perna, S. (2024). AI-Awareness: Fostering

Teachers’ Awareness of Artificial Intelligence Through

a Board Game. Lecture Notes in Computer Science (pp.

421–426). Springer Nature Switzerland.

https://doi.org/10.1007/978-3-031-78269-5_45

Sanders, E. B.-N., & Stappers, P. J. (2014). Probes, toolkits

and prototypes: Three approaches to making in

codesigning. CoDesign, 10(1), 5–14.

https://doi.org/10.1080/15710882.2014.888183

Saravanan, B., Kalaiyarasi, S., & Jeevarathinam, N. (2022).

Role of artificial intelligence in remote learning during

COVID-19 pandemic. Journal of Information

Technology and Digital World, 3(4), 307–319.

Stanford d.school. (n.d.). Design thinking bootleg.

https://dschool.stanford.edu/resources/design-

thinking-bootleg

Trendov, N. M., Varas, S., & Zeng, M. (2019). Digital

technologies in agriculture and rural areas. Briefing

Paper FAO.

http://www.fao.org/3/ca4887en/ca4887en.pdf

UNESCO. (2021). Education in a Post-COVID World:

Nine ideas for public action. Paris: UNESCO

Publishing.

van Amstel, F. M. C., Gatti, M. G., & Basques, F. A. (2012).

Design cards for creativity: Promoting knowledge

exchange through storytelling. Proceedings of the

Design Research Society Conference.

https://doi.org/10.1145/2371951.2371990

Vandenberg, J., et al. (2022). Promoting AI education for

rural middle grades students with digital game design.

Proceedings of the 54th ACM Technical Symposium on

Computer Science Education

, v.2, pp.

UNESCO. (2020). UNESCO-UNICEF-World Bank survey

on national education responses to COVID-19 school

closures – Key results (2nd iteration).

https://uis.unesco.org/en/news/unesco-unicef-world-

bank-survey-national-education-responses-covid-19-

school-closures-key-0

Wang, Y., & Lin, H. (2019). Research on the strategy of

balanced development of urban and rural education

based on modern information technology.

Wang, Y. (2020). Educational equity in the age of artificial

intelligence—Taking the construction of rural teachers

as an example. US-China Education Review, 10(4),

183–191.

Wölfel, C., & Merritt, T. (2013). Method card design

dimensions: A survey of card-based design tools.

Proceedings of the 6th International Conference on

Designing Pleasurable Products and Interfaces.

https://doi.org/10.1145/2513506.2513537

Dang, H. A., Oseni, G., Zezza, A., & Abanokova, K.

(2021). Impact of COVID-19 on learning: Evidence

from six sub-Saharan African countries.

Xiao, B., et al. (2022). Research on the model of balanced

development of compulsory education and difficult

problems of rural education based on a deep learning

algorithm. Wireless Communications & Mobile

Computing.

Yang, Y., & Zheng, Y. (2021). Research on balanced

development of urban and rural compulsory education

in the context of artificial intelligence. In 6th

International Conference on Education Reform and

Modern Management (ERMM 2021), Atlantis Press,

pp. 195–198.

Yousefi, S., & Nekoeimehr, M. (2020). Assessing the

susceptibility of schools to flood events in Iran.

Scientific Reports, 10(1).

Zhou, F. J. (2022). Methods to improve the efficiency of

rural physical education teaching resources allocation

and utilization in the context of artificial intelligence.

Computational Intelligence and Neuroscience.

CSEDU 2025 - 17th International Conference on Computer Supported Education

162