Identifying Rural Elementary Teachers’ Perception Challenges and Opportunities in Integrating Artificial Intelligence in Teaching Practices

Abstract

This research investigates the challenges and opportunities rural elementary teachers perceive in using AI as a pedagogical tool to support student learning in rural schools. Using a convergent parallel mixed methods approach, we analyzed the responses from 45 rural teachers who participated in professional development on AI integration in rural education. Through both closed-ended and open-ended survey responses, we employed an adaptation of the TPACK framework (I-TPACK) and the AI literacy framework proposed by UNESCO to identify the primary challenges and opportunities in utilizing AI for pedagogical purposes in rural education. The results highlight resource accessibility and teacher professional development as critical challenges and opportunities to reduce the digital divide in rural communities. Teachers see the inclusion of AI as an opportunity to personalize learning, reduce workload, and facilitate teaching in multigrade classrooms without perceiving it as a job threat. At the same time, they emphasize the need for technological and didactic resources aligned with the specific characteristics of their contexts, such as offline resources and adaptable AI curricula to address the prevalent issue of limited or absent internet connectivity in many rural schools.

1. Introduction

The impact and potential of artificial intelligence (AI) to transform teaching and learning practices is widely recognized [1,2]. AI technologies are being implemented to support student learning across multiple disciplines, multiple years, and teacher preparation [3]. In their work, Wang et al. [4] delineate four primary categories of AI application within education: adaptive learning and personalized tutoring; intelligent assessment and management; profiling and prediction; and emerging technologies, including educational robots, virtual reality (VR), and augmented reality (AR). These categories encompass a wide range of potential applications, each promising to improve educational outcomes significantly. However, several challenges remain. First, there is a lack of pedagogical foundations and support in AI education research [5,6] and a lack of consideration of teacher opinion in the AI design [7]. Second, there are biases regarding for whom and for what contexts the technology is developed, as it primarily focuses on urban and traditional teaching settings. Considering the already well-documented technological gap between rural and urban schools [8], designing AI around only urban contexts is exacerbating the gap and opportunities for rural students [9,10].

There is no doubt that AI has the potential to transform education, but the current state of AI is far from transforming teaching and learning practices effectively [11]. Several authors have pointed out that the lack of pedagogical rationale [5,6], poor pedagogical grounding [12,13], and misalignment between the educational community and technological development of AI are some of the major challenges to integrating AI successfully [14,15]. Considering that teachers’ beliefs, skills, and competencies play a critical role in technology integration in schools [16,17], the development of technology must carefully consider teachers’ perceptions, potential pedagogical usage, and specific needs [14,18]. Unfortunately, limited research investigates rural teachers’ opinions, perceptions, opportunities, and conceptualizations [19,20]. This research aims to contribute to filling this gap by analyzing rural teachers’ conceptions, challenges, professional development needs, and opportunities identified to integrating AI into their practices.

Access to, literacy in, and opportunities to use AI to support teaching and learning practices for all students are some worldwide priorities. “Viability and access to AI-learning opportunities for all learners is essential to prevent deepening the existing digital divide and avoid creating new disparities in education” [2]. This aspiration is supported by numerous UNESCO member countries and is recognized as a key factor in promoting educational equity and reducing existing gaps [21]. In this context, UNESCO [7] has urged States to implement collaborative actions that leverage the potential of artificial intelligence technologies to support the achievement of the 2030 Education Agenda while ensuring that their application in educational contexts is governed by the fundamental principles of inclusion and equity [21].

Unfortunately, this ideal is far from reality, especially when comparing Information and Communications Technology (ICT) access and literacy among students in rural schools versus their urban counterparts [8]. Despite the exponential growth of AI education research in recent years [3,22], the focus has been on traditional schools and students.

To address the lack of pedagogical foundations in AI integration and the increasing digitalization gap among rural students, this research explores the perceptions of 45 rural teachers from southern Chile regarding the challenges, needs, and opportunities to include AI literacy activities in their curriculum and use of AI to support their teaching practices. This research is framed within professional development workshops on AI literacy and classroom integration. These workshops began by defining traditional AI and the emerging trend of generative AI. For this research, we use “AI” to encompass both types of applications without making specific distinctions.

To analyze these perceptions, our research utilizes Celik’s [23] Intelligent Technological Pedagogical and Content Knowledge (I-TPACK) framework [23]. This framework is an adaptation of the Technological Pedagogical and Content Knowledge (TPACK) model developed by Thompson and Mishra [24], incorporating AI knowledge into the preexisting three dimensions.

Furthermore, we leverage the UNESCO AI literacy framework [7] to promote AI literacy as an objective. By using I-TPACK, which adds the dimension of AI to TPACK alongside the UNESCO framework, we acknowledge that the current advancements in generative AI are capable of generating new teaching practices that conceptualize AI in a dual role: as both a pedagogical tool and a legitimate member of a learning community.

The objective of this study is to investigate the opportunities and challenges that facilitate the integration of AI in rural contexts. To address this research, we explored rural teachers’ perceptions of including AI in their teaching practice, guided by the following research questions (RQs):

• RQ1: What opportunities and challenges do rural teachers perceive when including AI as part of their curriculum to promote AI literacy among their students?
• RQ2: How do rural teachers perceive the inclusion of AI in rural schools as a didactic tool for developing learning and as a tool to support educational management? Additionally, what are some of the benefits?

By answering RQ1, we aim to understand how to integrate AI literacy experiences into the rural curriculum. With RQ2, we expect to find guidance on strategies to develop AI tools that enhance classroom instruction and effectively support the overall teaching and learning experience.

2. Theoretical Framework

2.1. AI TPACK Framework

Shulman, in [25], developed the Pedagogical Content Knowledge (PCK) model to highlight the importance of transforming domain knowledge in teaching practices specific to this subject. Park and Oliver defined PCK as “teachers’ understanding and enactment of how to help a group of students understand specific subject matter using multiple instructional strategies, representations, and assessments while working within the contextual, cultural, and social limitations in the learning environment” [26] (p. 264).

PCK is operationalized under two dimensions: Pedagogical Knowledge (PK) and Content Knowledge (CK). Pedagogical Knowledge refers to the knowledge that instructors have about teaching methods and theories, classroom management, and knowledge that guides how to teach. Content Knowledge is the knowledge of a domain and the characteristics of this subject. Mishra and Koehler [27] proposed integrating a third dimension into the PCK model by adding Technological Knowledge (TK), leading to the Technological Pedagogical Content Knowledge (TPCK) model. This model was later adapted and renamed as Technological, Pedagogical, and Content Knowledge (TPACK) by Thompson and Mishra [24]. The TPACK model describes the knowledge required to develop specific teaching strategies using technology, create educational materials, and develop resources. It also explains how technological knowledge can help address some of the challenges faced by students [28].

TPACK is the most cited educational framework guiding research on technology integration in teaching and learning settings [24]. Furthermore, the TPACK framework is widely used to support teacher development programs for in-service teachers [29,30].

Despite the usefulness of TPACK, it does not explicitly consider the rise of AI in education. Celik [23] proposed an adaptation to measure teachers’ knowledge of the instructional use of AI. This model, called “Intelligent TPACK”, includes the following components: Intelligent-TK, the knowledge necessary to use AI tools with an understanding of their functionality; Intelligent-TPK, knowledge of the pedagogical possibilities offered by AI tools, such as personalized feedback and monitoring of student learning; Intelligent-TCK, knowledge of AI tools specific to each area of study and which are most suitable for promoting curricular learning; Intelligent-TPACK, teachers’ professional knowledge to select and use appropriate AI-based tools within a teaching strategy to achieve instructional objectives; and ethics, the analysis that teachers must carry out when making decisions about the use of AI-based tools. Our research is based on Celik’s model [23], which operationalizes these dimensions in the data collection instrument used.

The current literature in educational research includes proposals that aim to describe the knowledge teachers need to teach with AI (e.g., [31,32]). Lambera [32] presents one framework that guides teachers in developing the technical, pedagogical, and ethical skills necessary to integrate AI into education. This framework includes a taxonomy of AI applications linked to teaching and learning practices, fostering self-reflection and alignment with pedagogical objectives. However, it does not consider a specific distinction regarding the nature of the different types of knowledge the teacher requires, which could limit their capacity to address differentiated training needs according to the educational context.

Compared to other frameworks used to define the knowledge teachers need for teaching with AI (e.g., [24,31,32]), Intelligent TPACK offers a more specific and tangible approach to analyzing the integration of technology, pedagogy, and content in the context of AI. Intelligent TPACK provides a more concrete framework for analyzing the integration of technology, pedagogy, and content in the context of AI. This level of detail is particularly useful for identifying where to focus attention when exploring challenges and opportunities in rural contexts. For instance, the dimensions Intelligent-TK and Intelligent-TPK are fundamental for evaluating teachers’ initial technical and pedagogical knowledge, providing clear indicators to explore and understand their technological competencies and their capacity to integrate AI into teaching. This not only helps to diagnose training gaps but also to guide effective interventions in educational contexts with limited resources.

It is important to highlight that, although the aforementioned frameworks focus on the use of AI as a tool to promote learning, it is also essential that teachers promote AI literacy as an objective in itself [21]. This implies understanding what AI is and how it works, developing basic concepts and skills; using AI tools with knowledge of their limitations and potentialities; and being aware of the social and environmental impact that their use entails in diverse contexts. For this reason, this study incorporates the recommendations of UNESCO [7] to integrate AI literacy in the early education curriculum, exploring the perceptions of rural teachers regarding its inclusion in the classrooms.

2.2. Rural Education

Rural education is a global phenomenon that emerges as a solution to guarantee access to education in remote, isolated, or sparsely populated areas [33]. In countries like the United States and Chile, around 30% of educational establishments are considered rural [34,35]. In others, such as Colombia, rural schools represent 67% of the education system [36].

While rural schools operate within the broader national education system, their unique geographical, cultural, and socioeconomic contexts present distinctive challenges [37]. These contexts often require adaptations such as implementing multigrade teaching, where a single teacher instructs students across several grade levels [33]. Further challenges arise from the specific circumstances of rural communities. Schools located near indigenous reserves or migrant farmworker settlements frequently encounter issues related to poverty and student mobility [38]. In isolated regions, declining populations can lead to dwindling enrollment, jeopardizing the long-term viability of these schools. This issue is prevalent in many countries, such as the United States, Chile, Germany, and more generally in Africa and Latin American countries [38,39,40].

Moreover, geographical barriers bring significant obstacles. For example, in the southern-austral zone of Chile, some schools are located on islands accessible only by boat, and in some seasons, children cannot attend due to weather conditions disrupting transportation [41]. Being situated in geographically complex areas, such as islands, also adds challenges to access and classroom use, as rural areas often experience less access to and stability in internet connectivity [42].

Rural schools often face deficits in infrastructure and a lack of digital devices (e.g., computers, tablets), which hinders the integration of digital teaching practices and exacerbates the technological gap in these educational centers [43]. The challenges faced by rural schools are heterogeneous. While certain challenges are common, their specific manifestations often vary by region due to differing cultural, political, and economic factors.

Rural schools are much more than educational centers; they are pillars of the community, becoming essential social and cultural hubs [39]. The role of the rural teacher transcends teaching and learning, as they are indispensable references and support figures for meeting the diverse needs of the communities and their students. These needs can range from managing road improvements and water supply to community vaccination campaigns and support with children’s food and heating [44].

Furthermore, on many occasions, rural teachers provide economic support to families and students who lack the necessary resources for transportation and school supplies [38,45]. To facilitate students’ arrival at schools in these contexts, many rural teachers assume the responsibility of picking up students from their homes daily, which often means that some must leave their own homes and stay at their schools from Monday to Friday to achieve this [44].

Within this context, AI literacy, coupled with the use of AI tools and improvements in technological infrastructure (such as internet connectivity and digital devices), could potentially play a role in rural education. These advancements may offer avenues to mitigate the digital divide and provide support for rural teachers in addressing challenges such as multigrade classrooms, isolation, and limited resources. However, to ascertain the effectiveness and appropriateness of these tools, a collaborative approach involving rural educators is necessary. This research aligns with that need, seeking to identify the perceptions and potential barriers identified by rural teachers themselves, whose expertise on the heterogeneous challenges of their environment is crucial for developing contextually relevant solutions.

2.3. AI in Education and the Digital Gap in Rural Schools

Since Alan Turing introduced the idea of intelligent systems, various definitions have been proposed to define AI systems in education (e.g., [46,47,48]). In this study, we will use the definition by Díaz and Nussbaum: “AI in education as an educational technology capable of detecting patterns in existing or in vivo data and making automatic instructional decisions that are developed or implemented for pedagogical purposes to enhance the teaching and learning process” [14] (p. 2).

Given the wide range of applications of AI in education, several literature reviews have been conducted focusing on specific AI systems [49], specific domains [50] and applications [51], or specific educational levels (e.g., higher education [3] or K-12 [52]). For example, in different educational levels, the literature offers experiences, such as Bang et al. [53], who developed a math tutoring system to deliver exercises in early education. Similarly, Bonneton-Botté et al. [54] created an AI system to provide support for writing development by offering useful feedback and generating new exercises, targeting kindergarten students.

Indeed, extensive empirical studies of AI in K-12 have led Zafari et al. [22] to conduct a systematic literature review, while a similar effort has been made in higher education, as demonstrated by Zawacki-Richter et al. [13] with their systematic review focused exclusively on studies in this domain.

In specific subject domains, there appears to be a predominance of systems designed to support English as a second language (e.g., Yang et al., [55]) and mathematics education (e.g., [14,56]). However, several less traditional areas have also been explored. For instance, AI has been used to support music education [57] and nutrition education [58]. Even though the proliferation of AI found by those authors, there is a significant lack of alignment between AI design and educational practice [6,12].

Technological tools are essential, as they promote equal opportunities and facilitate digital literacy, a fundamental skill for 21st-century citizens [59,60]. However, despite the importance of technology use, many rural schools around the world still face gaps in access in, use of, and knowledge of technology (e.g., South America [61], South Africa [18], Europe, Guillén [62]). This digital gap creates a significant challenge to ensuring equitable educational opportunities for students in rural areas.

To address this gap, it is essential to ensure that rural schools have both access to the internet and the necessary digital literacy to effectively utilize new technologies [7,63]. Various educational policies have been promoted that seek to promote equality in the access to and use of technology [64,65]. These efforts are based on constitutional rights and ratified international treaties that seek to ensure that all students have the same learning opportunities [7]. Unfortunately, these policies do not seem to positively impact student learning in language, mathematics, science, or developing digital skills [64,66]. For instance, Li et al. [8] evaluated 5037 elementary school students in rural areas versus 5045 students in urban areas and found that rural students had statistically significantly lower ICT literacy, resilience, and online learning capabilities.

The teaching staff is a critical factor in the effective use of technology in rural schools. Teachers’ attitudes towards technology and their training to incorporate it into their practice are aspects that influence its use in the classroom [29]. Teachers that have a good attitude towards digital tools for administrative tasks might have difficulties incorporating them effectively into their classroom practices and do not always master all the technological skills they are expected to promote in their students [67].

Several proposals have been developed to prepare teachers to integrate artificial intelligence into their classroom practices. For example, Ding et al. [68] designed a professional development program to enhance AI literacy and integration strategies through pedagogical cases. This program combined direct instruction on fundamental AI concepts with discussions based on practical cases, allowing teachers to explore specific scenarios and applications in educational contexts. The results showed a significant increase in AI literacy, particularly in knowledge and understanding of this technology. This highlights the importance of combining direct instruction with case analysis to strengthen technical knowledge and the practical skills necessary for teaching.

Similarly, Zhao et al. [69] addressed the development of AI literacy among primary and secondary school teachers in China, focusing on four key dimensions: understanding fundamental AI concepts, applying AI tools in teaching, critically evaluating their pedagogical effectiveness, and reflecting on their ethical and social implications. Their findings highlight that the practical application of AI had a significantly positive impact on the other dimensions, underscoring its importance in teacher literacy.

The aforementioned studies demonstrate various approaches to addressing AI literacy among teachers. However, they also highlight significant barriers, such as limited technological knowledge, low awareness and acceptance of AI, and ethical concerns. These barriers are likely to be amplified in rural schools, which often face limitations in technological knowledge and access compared to their urban counterparts. Furthermore, factors such as attitudes toward AI, technological anxiety, preparedness, self-transcendence goals, and confidence in learning also influence the engagement and effectiveness of these initiatives [70,71]. The impact of these factors within rural settings has yet to be fully explored.

This scenario raises the need to promote actions that allow teachers to develop such skills.

3. Materials and Methods

3.1. Research Context

This research was conducted in the Los Lagos region, a southern city of Chile, a middle-income country and a leading educational nation in Latin America. Due to Chile’s natural geography, 83% of its territory is rural [37], encompassing approximately 3247 rural schools that serve over 35,550 students. Rural schools constitute about 30% of all schools in the nation. The southern part of the country hosts most of these rural schools, some of which are in remote areas with uni-docent and multiclass settings. Hautamaki [72] provide a description and visual context of some schools in the south of Chile.

Los Lagos, a rural city in southern Chile, has the second-highest number of rural public schools in Chile. Rural education in this region is primarily concentrated in elementary schools (75% of rural schools) and 63% of these schools operate with multigrade classrooms [37], where children of various ages and educational levels, with differences of up to 7 years, learn together [73]. Most rural schools are uni-teacher schools, meaning that one teacher manages the entire school, fulfilling all roles, including principal, administrator and teacher for the students.

Chile is in the early stages of focusing on AI, establishing policies, good practices, and regulations on its use. For example, in 2021, the government developed and launched the first policy on the use of AI [74], based on UNESCO guidelines [7]. These advancements have positioned Chile as a leader in data infrastructure, advanced human capital, research, and connectivity among its Latin American peers (see [75]).

Chilean teachers from rural schools were invited to a professional development workshop hosted by a local private university. Participation was voluntary and took place in December 2023, after the academic year had concluded. Teachers interested in participating in the workshop were previously informed about the objective of the study and its implications. At the beginning of the session, informed consent was obtained from all individuals interested in participating, in accordance with the Declaration of Helsinki. The study was approved by the Ethics and Bioethics Committee of the Universidad Austral de Chile on 22 April 2022. To ensure data privacy, the first author anonymized participants’ personal information and assigned them a numerical code for identification.

The workshop focused on AI in rural education. This four-hour workshop was divided into two main blocks. The first block aimed to introduce rural teachers to artificial intelligence, giving them a general understanding of its operation and highlighting its presence in everyday life. This introduction was carried out through an inaugural conference that sought to demystify AI and show its influence on various aspects of our daily lives.

The second block focused on practical activities, where teachers explored the advantages and limitations of AI and connected AI literacy with the national curriculum. To begin, teachers worked in small groups with two object detection applications developed specifically for the workshop, as shown in Figure 1. These applications, which functioned offline, allowed for the identification of common classroom objects such as chairs, tables, and people. The teachers freely explored the applications, evaluating their functionalities and limitations.

Subsequently, the session delved into the axes of AI literacy proposed by UNESCO [7], using practical examples to facilitate understanding. The teachers were invited to identify connections between these axes and the national curriculum and propose how to implement these activities in their own educational contexts. For example, within the axis of ethics and social impact, a Deepfake created from an image was presented, allowing teachers to reflect on the social impact of these technologies and discuss which learning objectives from the national curriculum could be addressed through this activity.

3.2. Participants

The study considered a non-probabilistic convenience sample. Rural teachers from different communes in the Los Lagos region were voluntarily invited to participate in the study.

The sample consisted of 45 rural teachers, with a detailed demographic breakdown provided in

Table 1. Teacher demographics.

Characteristic	Value
Sample size	45
Gender identified	41
Women	65.9%
Men	34.1%
Age range	23–66 years
Average age	43.4 years (sd: 11.46 years)
Average teaching experience	14.43 years (sd: 10.94 years)

. The teachers had diverse backgrounds, reflected in their wide age range and varying years of experience in the classroom. The gender distribution of the teacher sample in this study is representative of the K-12 teacher population in the Chilean educational system [76]. For example, between 2002 and 2021, the proportion of female teachers in Chile ranged from 70% to 73% [76], in both rural and urban areas. This is comparable to other countries such as the USA, where 77% of teachers identified as women in 2020–21, according to the National Center for Education Statistics [77].

Notably, more than half of the teachers (54.5%) worked in single-teacher classrooms, where they were responsible for teaching up to six grade levels simultaneously, where 18.2% worked in two-teacher schools, teaching up to three grade levels at the same time, and 27.3% were teachers in traditional rural schools, teaching one grade level at a time. Regarding connectivity, 11.6% of the participants reported not having connectivity at their school, 51.2% stated that they had an intermittent connection, and 37.2% reported having good connectivity.

While the sample size is relatively small, it is sufficiently large to explore the perceptions and experiences of rural teachers with AI integration in classrooms. Additionally, we believe the sample adequately captures the diversity of rural school contexts, including variations in teaching load, connectivity issues, and school organization. This diversity ensures that the findings are relevant to a wide range of rural education settings.

Regarding the qualitative data, we carefully considered saturation in the analysis. Given the nature of the study and the depth of responses in the open-ended questions, we reached theoretical saturation in the qualitative data, with no new significant themes emerging after a certain point. This supports the reliability of our findings and ensures that the sample is sufficiently representative of the experiences and perceptions of rural teachers regarding AI integration.

3.3. Instruments

This section details the instruments used to gather data for this study, which explores rural teachers’ perceptions of AI and its integration into the educational curriculum. The research employed a mixed-methods approach, utilizing qualitative and quantitative data collected through a survey instrument. The survey was designed using the theoretical framework of TPACK as outlined by Celik [23] and the AI literacy framework proposed by UNESCO [7]. The research was guided by two central research questions: RQ1, focusing on integrating and promoting AI literacy in the rural education curriculum, and RQ2, seeking to evaluate rural teachers’ perceptions of AI and the potential pedagogical uses they conceptualize.

To address RQ2, we employed two dimensions of Celik’s [23] Intelligent Technological Knowledge: one to measure teachers’ familiarity with AI-based tools (Intelligent-TK) and the other to measure their pedagogical understanding of AI (Intelligent-TPK). These dimensions were evaluated in the first part of the survey using closed-ended items with a Likert-type scale. Our decision to focus specifically on Intelligent-TK and Intelligent-TPK stems from the rural context of this study, where teachers often face limitations in technological infrastructure and prior exposure to AI tools [7]. Other dimensions, such as Intelligent-TCK, Intelligent-TPACK, and Ethics, were deemed less relevant for this initial assessment as they require access to specialized tools or presuppose prior AI experience, often lacking in these settings.

The Intelligent Technological Knowledge dimension was operationalized to explore teachers’ level of familiarity with the technical capabilities of AI-based tools, and Intelligent Technological Pedagogical Knowledge to explore the pedagogical possibilities of AI-based tools that teachers envision. These dimensions were assessed using 31 closed-ended items with a Likert-type scale. The Likert items were organized into three subsections within this first part of the survey. The first subsection addressed perceptions of AI, with statements such as, “When I hear the term artificial intelligence, I associate it with machines with a certain level of intelligence that will replace humans in multiple tasks; data models and training; facial and voice recognition applications; and ChatGPT, among others”. The second subsection explored the opportunities and challenges of using AI tools to manage the educational process. This section included statements like, “The main challenge of using artificial intelligence tools for managing the educational process is the extra time required for their use”. The third subsection explored the opportunities and challenges of using AI tools as didactic resources to promote curricular learning. This section included items such as, “The main opportunities offered by using AI tools as didactic resources to promote curricular learning is to motivate learning”.

The second part of the survey centered on exploring how to integrate and promote AI literacy in the rural education curriculum (RQ1). This included aspects like the use of AI tools, the functioning of this technology, and the ethical considerations associated with its implementation. To guide this analysis, we utilized UNESCO’s [7] recommendations, which focus on three areas for AI literacy in early education curricula: (i) understanding how AI works and its capabilities, (ii) understanding and using AI techniques and technologies, and (iii) ethics and social impact. These areas were evaluated through closed- and open-ended questions in this second part of the survey. Specifically, a closed-ended section with 18 Likert-scale items was included, with questions such as, “The main opportunities provided by teaching what artificial intelligence is, how it works, the tools it uses, and its ethical and social implications in rural education are the development of necessary skills for 21st-century citizenship”. These perceptions were further explored through open-ended questions in the final section of the survey, with prompts such as “As a rural teacher, what challenges do you visualize for the integration of AI literacy with subjects in the national curriculum?”.

3.4. Data Analysis

Quantitative and qualitative data from the survey were integrated to address both research questions. Qualitative responses from the survey’s final section enriched the analysis of rural teachers’ perceptions of AI as a didactic tool and for educational management support. Although no open-ended questions were specifically dedicated to this aspect (RQ2), teachers addressed it alongside challenges related to AI literacy integration (RQ1) in their open-ended responses. These qualitative data provided deeper insights into perceived benefits, specific experiences, and rural context considerations, complementing the quantitative analysis.

Qualitative data from the open-ended questions were analyzed using a hybrid approach—a combination of inductive and deductive thematic analysis from Fereday and Muir-Cochrane [78]—guided by the adaptation stages of educational qualitative research outlined by Xu and Zammit [79]. This approach was built upon the six stages of thematic analysis developed by Braun and Clarke [80]. Initially, we extracted data from the survey using the pre-generated themes (e.g., challenges and opportunities) that were foundational to the development of our research and research questions. Since our instrument was specifically designed to identify these dimensions, the data segmentation process was straightforward and conducted by the first author of this paper.

After segmenting the data into discrete parts, an inductive thematic analysis was conducted, beginning with open coding of each discrete response to the questions. Two members of the research team carefully read each response and summarized them with a word or short phrase using an in vivo coding process as described by Guest et al. [81]: “coding for words or phrases within the text” (p. 35) For example, one teacher mentioned, “Rural areas suffer from a lack of connectivity and technological resources” (participant #10), which we coded as “resource limitations”.

Both researchers collaboratively coded all the data using in vivo coding. We created an Excel spreadsheet with a column containing all the keywords extracted from the in vivo coding, summarizing each response. Additionally, we developed specific definitions for each code and included quotations from participant responses to illustrate the codes.

Next, axial coding was conducted. We carefully reviewed all the emerging codes to identify similarities and differences among them. Categories were created to group codes with similarities. For instance, codes like “resource limitations” and “lack of infrastructure” were grouped into the category “Resource Constraints”. As researchers, we are also mindful of how our backgrounds may influence this study.

Since we conducted this research in a Spanish-speaking country, we performed all analyses using data preserved in Spanish. All authors are native Spanish speakers. For direct quotations in this manuscript, we employed a two-stage translation process. This two-stage process involved utilizing AI tools for initial consistency. Subsequently, the team members fluent in English meticulously reviewed and edited the translations. This approach ensured consistency in treating grammatically incorrect phrases while maintaining clarity and readability for an English-speaking audience.

For the quantitative analysis of the closed-ended items, a descriptive frequency analysis and association analysis using the chi-square statistic were conducted using Jamovi, v.2.3.18. Specifically, we aimed to reflect the most frequent conceptualizations, concerns, and perceptions of opportunities and challenges. We focused on identifying the most common perceptions of opportunities and challenges. Then, we evaluated whether these perceptions were significantly correlated with the practical implementation of activities in the classroom, highlighting those with the greatest relevance for practical use in the rural education environment.

3.5. Data Integration

Quantitative and qualitative data, both collected through the survey, were integrated into the analysis as part of a data triangulation technique [82], specifically using a triangulation design convergence model approach [83]. The research questions served as the main themes for triangulation.

For example, in the quantitative section, participants reported how much prior experience they had with AI and the frequency with which they used AI in their teaching practices before the professional development training. In the qualitative section, participants discussed the challenges they encountered when attempting to use AI. By combining these data, we analyzed how teachers with varying levels of AI usage experience related to the challenges they reported and the frequency of those challenges. Additionally, when participants were asked about future opportunities to use AI to support teaching practices, we analyzed the quantitative data regarding their prior experiences and integrated it with the qualitative responses to gain a more nuanced understanding of their perspectives.

By employing this approach, which combines qualitative and quantitative data, we aim to achieve a better understanding of the research problems this study seeks to address.

4. Results

4.1. RQ1. Promoting AI Literacy in Rural Education

We investigated the opportunities and challenges for AI literacy in rural schools, considering the specific characteristics of these contexts. To do this, we analyzed the teachers’ closed-ended responses and their open-ended responses.

4.1.1. Challenges

Rural teachers identified a lack of professional development as the most significant challenge in teaching AI. With an average rating of 4.7—the highest across all items on a scale of 1–5—teachers strongly agreed on the need to improve their professional development in this area. In particular, three challenges received unanimous agreement from participants, with 100% positive ratings: the lack of teacher preparation to teach AI, the integration of AI into the school curriculum of rural schools, and the relevance of AI for the future of students. Additional challenges with a high consensus among participants include the lack of resources that do not require internet connection and are accessible to rural schools (97.7%), the lack of didactic guidelines for teaching artificial intelligence in diverse rural contexts (97.7%), and the increasing digital divide experienced by rural schools (93.2%). All participants at least agreed that more professional development is necessary. This finding aligns with the qualitative analysis, where “professional development” emerged as the most frequent theme, with 33 coded instances among the 49 responses. It is interesting to note that the responses associated with professional development explicitly connected elements of the framework of TPACK as outlined by Celik [23]: Intelligent-TK (recognizing the need to know basic AI concepts) and Intelligent-TPK (recognizing the need to learn pedagogical strategies to incorporate AI strategies in teaching). For example, some responses highlighting this need include the following:

“Ongoing training to learn strategies for the uses and scope of AI literacy in the curriculum”. (Participant #1)

“A challenge is to learn the basic concepts to be able to explain them to our students and put them into practice” (Participant #22)

However, 13 out of the 33 respondents did not provide specific details about which areas of AI require more attention and lacked further elaboration. For instance, some responses stated

“Teacher improvement in AI”. (Participant #4)

“Train teachers”. (Participant #38)

The remaining 20 respondents provided more specific subcategories, such as knowledge of AI (e.g., Participant #1), pedagogical content knowledge related to AI (e.g., Participant #22), and AI for interdisciplinary development, as illustrated by the following:

“Learning to integrate AI into all subjects of the curriculum”. (Participant #27)

The second most highlighted challenge in the quantitative survey was the “Lack of didactic guidelines for teaching AI in diverse rural contexts (e.g., single-teacher or two-teacher schools, traditional rural schools)”, which received an average rating of 4.66 (on a scale of 1–5), with 100% of the teachers agreeing on this being a significant challenge. This issue was mentioned in the open-ended responses, but it had a low explicit frequency. For example, one respondent connected the need for professional development to the specific challenges faced by single-teacher schools:

“Teacher training (because being a single-teacher school makes learning activities [courses, workshops, etc.] more complex)”. (Participant #49)

In general, respondents mentioned their challenges in broad terms, without specifically attributing them to their unique contexts.

The third challenge, “Increasing the digital divide in rural schools” (93.2%), was also agreed upon by 100% of the teachers as a major issue in their context. From the open-ended questions, we identified that this digital divide is framed within the context of resources and technical–administrative conditions. Specifically, 29.5% of the participants mentioned technical–administrative conditions, and 30.3% cited a lack of available resources. These challenges are related to the conditions under which rural teachers operate, such as certification, training opportunities, and support from key actors in the educational administration system. For example, one teacher commented

“The tools that should be available and provided for the benefit of all, ensuring more technological resources and becoming a relevant factor considered in the management of the funder, while respecting allocated timeframes”. (Participant #2)

Another teacher noted

“Increasing hours in technology at the expense of other subjects”. (Participant #2)

And another teacher also shared

“The main challenge is the flexibility of the national curriculum to make it adaptable to the rural context”. (Participant #12)

These comments highlight the need for the national curriculum to be flexible enough to allow the implementation of AI. Teachers mentioned that this support must come directly from the Ministry of Education:

“The Ministry of Education’s commitment to the transformation of new learning approaches”. (Participant #3)

In terms of resources, teachers highlighted internet connectivity and the lack of technological devices as challenges that create additional barriers in rural schools. For instance, one respondent mentioned

“Each school is unique and should be approached according to its own characteristics. Moreover, we lack adequate internet access for work, and not everyone has access to technological devices”. (Participant #18)

An example of the need for satellite connections was also provided:

“The challenge lies in providing satellite connectivity to rural areas”. (Participant #48)

Another teacher mentioned

“The challenges I see are the lack of technological tools and internet accessibility”. (Participant #43)

This aligns with the fourth highest-rated challenge in the closed-ended questions: “Lack of resources that do not require internet connection and are accessible to rural schools”, with a rating of 4.5. Teachers involved in more robotics activities tend to view this challenge as less relevant, with statistical evidence (${\chi }^2$ = 16.119, p = 0.013) supporting this tendency. This could mean that teachers who actively use robotics are more accustomed to working with offline resources or tools, possibly due to the nature of robotics projects, which often do not rely heavily on internet connectivity. These teachers are also more likely to disagree with the statement “I do not feel prepared to install it”, indicating that they feel more capable of implementing these technologies in their classrooms. This result (${\chi }^2$ = 21.596, p = 0.042) suggests that teachers who are more familiar with robotics may feel more confident in handling technological installation and integration.

On the other hand, the results for teachers who engage more in AI activities are less definitive. While no statistically significant findings were observed for any specific items, there was a trend (${\chi }^2$ = 12.311, p = 0.055) suggesting that teachers who implement more AI activities also consider the lack of offline resources to be a less relevant challenge. This may imply that AI activities, although requiring internet access, might also involve tools that can be adapted to offline settings. For AI activities to be successfully integrated into rural schools, they must be capable of functioning offline, ideally with minimal or no internet connection.

Finally, other challenges identified by participants in the closed-ended questions include ethical considerations (4.1 average agreement), class time, and the relevance of AI to the future. These were also recognized as mid-to-high agreement challenges by the participants.

4.1.2. Opportunities

A majority (95.4%) of rural teachers recognize the potential of AI and the importance of incorporating AI literacy in rural schools, as indicated in the responses from the survey’s second part. There is a strong consensus among these teachers on the value of integrating AI literacy into rural education to help students “become aware of the risks in the digital world and the impact of false information”. Interestingly, only one participant disagreed with this statement in the Likert-scale responses. The results of the survey suggest that teachers who most considered that AI literacy creates an opportunity to understand the presence and influence of AI in the world in which students live tended to generate more activities related to programming, ${\chi }^2$ = 15.746, p = 0.046. In the case of robotics activities, they were associated with the perception that AI literacy in rural contexts creates an opportunity to raise awareness about risky behaviors in the digital world and the impact of false information, ${\chi }^2$ = 13.105, p = 0.041. In other words, teachers who are mindful of the risks and ethical concerns surrounding AI seem to be more inclined to focus on robotics activities, perhaps because they perceive robotics as a practical and engaging way to teach students about technology’s role in society while promoting responsible use.

We examined their responses to the open-ended question about opportunities, and one teacher stated

“I see an opportunity to incorporate new tools into the classroom. Furthermore, students and teachers will be able to access technology that they may not have at home. This also presents opportunities for co-teaching by integrating subjects”. (Participant #42)

Even though Participant #42 disagreed with the idea that AI presents opportunities to address risks and dangers in the digital realm, they still saw its potential as an opportunity, particularly given the role rural schools often play in their communities. Schools in rural areas frequently provide access to technology and resources that students might not have at home. Additionally, this participant highlighted the potential for interdisciplinary integration—a topic of significant interest.

The theme of interdisciplinarity was the most common one mentioned by participants in their open-ended responses. Teachers view AI tools as pedagogical instruments that can help them build transdisciplinary connections. For example, some of the responses included the following:

“To make use of the cross-curricular nature of this tool with different subjects”. (Participant #8)

“It is about providing tools to work with in all subjects, and it is an opportunity that must be seized” (Participant #16).

It is likely that teachers emphasize the capacity for interdisciplinarity because, in their rural context, they often have to teach multiple subjects across different grade levels within the same classroom. Consequently, rural teachers naturally adopt teaching practices that emphasize transdisciplinarity and progressive learning. This aligns with the traditional research focus on AI in education, which often emphasizes context-specific approaches and supports our initial expectation to not exclusively focus on the dimension of PCK.

The second most prominent theme, emerging from both qualitative and quantitative data, is that AI has the potential to reduce the current gap between rural and urban education in terms of technology access. In the quantitative survey, this item received a high agreement rating of 4.6, and 22.4% of the codes in the qualitative analysis were assigned to this category. Some of the highlighted responses include the following:

“It breaks the rural-urban divide, creates awareness among students about AI”. (Participant #9)

“It provides an opportunity to reduce the gap between rural students and those in traditional schools”. (Participant #10)

“To reduce the gap that currently exists between rural and urban students”. (Participant #29)

Indeed, teachers are acutely aware of the technological access and opportunity gaps between rural students and their peers in urban schools.

4.2. RQ2. Using AI to Support Teaching and Learning in Rural Schools

As previously mentioned, the survey did not incorporate open-ended questions addressing RQ2. However, in responding to open-ended questions about the challenges of AI literacy integration (RQ1), teachers also touched upon their perceptions of AI’s pedagogical uses (RQ2). These insights were incorporated to enhance the analysis of RQ2.

The most common theme that emerged regarding the implementation of AI in rural environments is its importance for the future. Many teachers emphasized that AI is an inevitable part of the future, and preparing students to understand and use AI is crucial. They believe that AI can enhance students’ skills relevant to the 21st century. For example,

“I believe that including AI in current teaching is necessary because our students were born in the digital age”. (Participant #28)

“As teachers, we cannot be oblivious to new knowledge, and we must prepare students for these new technologies, which can be a tool for their future”. (Participant #22)

“It should be included as a basis for developing life skills since AI is currently present in almost everything we do. All of this, as long as a real awareness of ethical and social norms is achieved to avoid misuse as much as possible”. (Participant #20)

These comments highlight how teachers are consistently aware of the need to prepare students for future success.

Three responses also emphasized that AI could improve student motivation through innovative teaching practices:

“I think it is positive, as it motivates students”. (Participant #9)

“Artificial intelligence can be included through various activities, on a computer or tablet, so that students can have greater motivation since it is very entertaining for them to work with these tools”. (Participant #17)

“Technology generates greater motivation for students”. (Participant #45)

These qualitative responses are consistent with the survey data, where nearly all teachers agreed that AI enhances student motivation. The 4.7 rating is supported by qualitative evidence that teachers recognize AI as a tool to foster engagement, particularly through its interactive and entertaining nature. This overlap demonstrates how teachers’ detailed reflections provide depth to the numerical survey results, offering a richer understanding of their perspectives on AI’s potential to motivate students. Only one teacher (Participant #34) did not agree or strongly agree with using AI for this purpose. This participant, a self-identified female who teaches a single-teacher class with intermittent internet access, has never actually implemented AI in her teaching practices. In contrast, the other 44 teachers were very optimistic about AI’s potential to enhance student motivation and behavior. Her divergent perspective contrasts with the overwhelmingly positive qualitative and quantitative responses from other participants, further emphasizing the importance of practical exposure and adequate resources to ensure more consistent adoption. Given that various studies have reported challenges with student behavior and attention, especially post-COVID, AI could play a significant role in supporting student motivation.

Several teachers advocated for integrating AI education, not only in Technological Education but across various subjects. They believe AI concepts should be taught in a way that intersects with other areas of the curriculum to provide a more holistic understanding. Some common responses included the following:

“Transversally with the curriculum between subjects Project-Based Learning establishes an improvement in collaborative and cooperative work among peers, including AI in the improvement of strategies or using it in the different units” (Participant #1)

“AI works transversally for all subjects”. (Participant #10)

“The teaching of AI can be included transversally in different subjects, with activities of interest to students”. (Participant #11)

“AI as an opportunity to promote learning in different subjects, both interdisciplinary and intradisciplinary, in addition to taking advantage of AI in the daily life of the educational community”. (Participant #15)

Unfortunately, only a few responses provided specific teaching practices for AI integration, with only three respondents highlighting the use of project-based or problem-based learning.

A significant number of teachers also emphasized the importance of teaching the ethical and social impacts of AI. They stressed that while AI can offer many benefits, it also poses risks such as job displacement and ethical dilemmas. For instance, one teacher commented

“I think it is necessary and that it should be established now because our children handle technology without proper guidance, even without anyone controlling it” (Participant #19)

The main concern is that AI is already present in students’ lives, and the fact that they are using technology without proper guidance is problematic. This was also highlighted in the survey.

Teachers overwhelmingly agree that the focus should not be on the dangers of AI but on teaching students how to use it responsibly and effectively. One teacher expressed this sentiment well:

“Today, guided and controlled approach to technologies is fundamental. Ethics, morals, empathy… should be the primary objective… AI at the service of the human being”. (Participant #3)

“To dispel myths, that machines will not replace us and that we will be the ones who get the best out of them for the benefit of our students”. (Participant #40)

This opinion aligns with the principles of Human-Centered AI [84] and the adapted concept of Pedagogical Human AI [14], where AI should be designed and implemented to serve humans rather than replace them.

Teachers identified several specific applications and benefits of AI literacy in their context, including using AI to reduce their workload, organize student information, improve teaching practices, and motivate students. For instance, here are some examples:

“Support in school management. Improve our pedagogical practices. Optimize teaching time and work”. (Participant #5)

“Opportunity to optimize time in carrying out different courses”. (Participant #32)

These results from the open-ended questions align with the quantitative findings from the survey. Furthermore, all teachers agreed that AI could support their teaching practices by assisting in task design and facilitating the creation of resources. These two areas received the highest ratings, with scores of 4.6 and 4.5, respectively. Indeed, all the rural teachers surveyed either agreed or strongly agreed that these were the primary ways AI could support them. However, they also emphasized that they did not feel adequately prepared to use these tools, particularly when it came to completing their specific tasks.

Additionally, with a score of 4.40, teachers acknowledged that they had not received sufficient preparation to implement AI-based teaching practices, mainly due to a lack of professional development opportunities. Furthermore, with a score of 4.36, teachers highlighted the need for tools that do not rely on internet connectivity, which is a significant issue in rural contexts.

5. Discussion

To address the challenges and gaps in access to and usage of technology within rural educational contexts [85,86], this research explores the perceptions of opportunities and challenges related to integrating AI literacy and AI into the pedagogical practices of 45 rural teachers from the southern region of Chile.

Our findings suggest two key aspects that, according to the teachers’ perceptions, represent both a challenge and an opportunity for the inclusion of both AI approaches in rural schools: (i) accessible resources for rural schools and (ii) teacher professional development. These aspects can aggravate or mitigate the digital divide experienced by many rural educational communities [63]. The COVID-19 pandemic highlighted that the digital divide continues to be a significant challenge for many rural schools worldwide [7]. In the case of AI, the digital divide involves difficulties in accessing existing AI technologies such as ChatGPT, virtual assistants like Alexa or Google, and intelligent tutors, among others [12], as well as in understanding how AI works, its limitations, and developing an understanding of how to use this technology ethically and responsibly [87].

The findings of this study show that rural teachers are aware of the digital divide that many rural communities still experience, associated with both access to resources and the knowledge necessary to use them. The implementation of AI in rural areas requires careful consideration of the unique challenges these settings present. Limited resources, unreliable connectivity, and the need for affordability necessitate AI solutions that can operate effectively on low-cost devices like tablets and low-end notebooks, often the only available technology in these regions. Additionally, they involve access to teaching materials that consider the heterogeneity of rural contexts and support their implementation in the classroom (teaching guidelines for multigrade schools and differentiated lesson plans, among others). The development of resources accessible to all schools is one of the critical challenges for the inclusion of AI in education and an imperative for international organizations such as UNESCO [7]. The latter seeks to ensure that the development of this new technology is an opportunity to develop more just and responsible societies, not a reason to generate inequality and exacerbate existing educational gaps.

To address these technological challenges, a focus on developing open-source, offline-capable AI models and algorithms is crucial. These solutions must be optimized for minimal computational requirements, ensuring they can function seamlessly in environments with limited internet access and processing power. This technological shift must be accompanied by educational initiatives that empower individuals in rural communities to build and deploy AI-powered products and solutions tailored to their specific needs. By prioritizing accessibility and sustainability, we can unlock the transformative potential of AI in rural settings, fostering economic growth, improving healthcare outcomes, and enhancing overall quality of life.

Our results show that rural teachers are aligned with this vision and see AI as a tool to help eliminate the gaps historically presented in rural schools. Furthermore, teachers recognize that using AI technologies and AI literacy in their students offers opportunities to implement motivating and relevant experiences that allow for developing 21st-century skills, addressing the heterogeneity of the rural classroom. The literature has shown rural teachers’ difficulties in managing several curricula simultaneously and designing and implementing activities for several grades simultaneously, as well as the lack of training and teaching materials that support such a purpose [14,88]. In this sense, AI tools are perceived as an opportunity to support the creation of materials that adapt to the particularities of each student, as well as to promote learning autonomously or in a guided way, in a more personalized manner, and to motivate learning.

Rural education faces unique challenges, such as teaching in multigrade classrooms, geographic isolation, scarcity of resources, and limited technological connectivity, all widening the educational and technological gap compared to urban areas [7,66]. While these difficulties persist, recent advances in artificial intelligence (AI), such as foundational models, offer promising opportunities to address these issues. Integrating AI into the curriculum of rural communities, taking into account the particularities of their context, can facilitate effective pedagogical strategies to manage the diversity of ages and learning levels in multigrade classrooms. Adaptive learning tools and intelligent tutors could support the work of teachers, optimizing their time and allowing for more personalized attention for students experiencing difficulties. However, there is also a risk that it could reinforce pre-existing inequities if not implemented with an inclusive and equitable approach. For example, the lack of access to AI tools in rural contexts could limit the development of essential technological competencies in students, reducing their ability to integrate effectively into the workforce or restricting their possibilities of accessing careers related to technology and other fields. The latter would not only perpetuate inequalities but also compromise the transformative potential of AI in reducing educational gaps and fostering sustainable development [7].

AI literacy is perceived as an opportunity to prepare students with the necessary skills for the 21st century and to integrate the teaching of multiple subjects. Future research could address this gap and explore what characteristics AI curricula for rural schools should consider, determine how the content areas of AI literacy could be linked to the existing curriculum, and explore their effects on rural students. In this context, involving local educators in the design process is essential, allowing the content to be tailored to their contexts’ specific needs and realities. Furthermore, it is crucial to integrate culturally relevant examples and practical applications that foster a connection with students, increasing their interest and engagement. These actions would ensure a more inclusive curriculum and a more effective one for promoting the understanding and responsible use of AI [7,21].

Professional development emerges as a critical and urgent action for the successful integration of AI in rural schools, which aligns with findings from other studies [86,89]. Teacher participation in training experiences, programs, or activities aimed at improving or acquiring new knowledge and skills and changing attitudes to enhance student learning outcomes is highlighted as essential [90,91]. The teachers participating in this study recognize the need for training to use this new technology in the classroom, and they raise different aspects that should be considered for this purpose. The first of these has to do with the technical–administrative conditions of teaching. They see the need for certification as a form of recognition for the time dedicated to continuing education and its recognition in the teaching career. They also highlight the allocation of actual time to participate in training instances. Similar studies have reported time constraints as representing an obstacle for teachers to carry out sustained training actions over time [88]. Future studies could explore how the allocation of actual time to participate in training impacts teacher motivation, persistence in continuing education programs, and the development of knowledge and skills.

One particularly interesting finding is that none of the participants expressed concerns about AI potentially “replacing their jobs”. Given the complexity of rural education roles, where teachers often feel overwhelmed with the multitude of tasks—many of which extend beyond teaching and learning activities—they view AI as an opportunity to reduce their workload and facilitate their practice. They do not perceive AI as a threat to their job security.

Finally, a line of research worth further exploration is the conceptualization of AI as a tool to foster interdisciplinary practices, particularly in the context of rural schools where multigrade settings are common. Teachers are interested in implementing practices that span multiple disciplines and grades. AI tools designed from a constructivist perspective, capable of identifying and proposing personalized learning experiences, seem highly promising for addressing the challenges of effective learning in the highly diverse environments typical of rural settings.

6. Limitations

A limitation of this study is the sample size of 45 participants. However, it is important to note that in the context of rural education, this sample size is quite substantial. For instance, our sample size is three times larger than that of Martin [92], three times larger than the study of ICT experiences among rural teachers in Nepal by Rana et al. [93], and four times more than the in-service teachers researched by Xu and Stefaniak [94], to list a few. However, we recognize that the small sample size may impact the statistical power, diversity, and representativeness of our findings. Future research could address this limitation by including a larger, more diverse sample to improve the generalizability of the results.

Secondly, all participants in this study are from the same region within one country. While this may limit the generalizability of our findings, there are shared similarities between our study and other rural education studies in different regions. For example, our findings are consistent with those from similar studies in South America (e.g., [85]) and Asia [29], suggesting that some of our conclusions may be applicable to rural education contexts in other parts of the world. However, we acknowledge the importance of considering regional variations, such as cultural and policy differences, in future studies to further validate these findings.

Thirdly, using a survey-type instrument, while effective for gathering general perceptions, could limit the depth and breadth of participants’ experiences. The data collected are limited to responses from a predefined set of questions, which may not fully capture the nuances of teachers’ perceptions and challenges. Future research could address this limitation by incorporating individual interviews, allowing for a deeper exploration of teachers’ perspectives and providing more comprehensive insights into their experiences with AI tools in rural education.

7. Conclusions

The inclusion of AI in education holds immense potential to uplift rural teachers in their mission to provide quality learning in highly diverse and heterogeneous classrooms. While it may exacerbate the digital gap these communities already grapple with, the potential benefits far outweigh the challenges. To ensure rural schools can harness this technology without being left behind, it is imperative to develop specific resources that cater to the unique needs of these contexts. This study offers valuable insights into the opportunities and challenges perceived by 45 rural teachers regarding the use of AI for pedagogical purposes, shedding light on how teachers envision its educational integration in rural areas. This study has identified two key aspects teachers perceive as both challenges and opportunities in integrating AI into rural classrooms: resource development and teacher professional development.

The first is resource development, a crucial factor for including AI. Teachers report a lack of technological devices and internet access as a significant barrier. However, they see an opportunity to develop offline devices that allow for personalized learning experiences adapted to the characteristics of their students, thus facilitating their educational work. In addition, resources for the pedagogical use of AI in rural contexts must include technology and specific didactic guidance, such as lesson plans and practical examples that can be applied in multigrade classrooms, where a teacher must simultaneously teach students of various levels. This task is particularly challenging, as existing guides on education do not usually consider this teaching modality.

The implications for educational practice underscore the urgent need to develop platforms or tools that function without an internet connection, as well as the creation of teaching materials tailored to multigrade teaching in rural areas. Teachers require resources that not only grant them access to technology but also offer practical pedagogical guides that they can implement in highly diverse environments with technological limitations. To address these challenges, partnerships between governments, Non-Governmental Organizations, and tech companies are essential to fund, develop, and distribute these tools. Furthermore, establishing funding mechanisms to ensure the sustainability of these resources over time will be critical. This approach not only mitigates the impact of the digital divide but also promotes the inclusive use of AI in these contexts.

The second aspect is teacher professional development. To fully harness AI’s potential, rural teachers need ongoing training opportunities that allow them to understand, adapt, and effectively apply these technologies in their specific contexts. Training in digital skills and the design of appropriate pedagogical activities for rural environments are essential to ensuring the successful integration of AI in these schools, as well as fostering AI literacy among teachers to promote awareness of potential biases in AI tools and ensure their ethical use, especially in communities that have been marginalized in terms of access to technology. To make this a reality, policymakers must prioritize the development of professional development programs and ensure accessibility for rural teachers. These programs should include interdisciplinary lesson plans and strategies tailored to rural contexts, enabling teachers to use AI to improve learning in classrooms with limited resources and diverse student needs. Collaborative efforts among stakeholders can ensure teachers have sustained access to these training programs, promoting long-term success.

The implications for educational practice here highlight the need to develop teacher training programs focused on the pedagogical use of AI, which not only teach how to manage the technology but also train teachers in strategies that leverage AI to improve learning in classrooms with students of different levels and with limited resources. This includes developing lesson plans that promote learning through interdisciplinary approaches, such as integrating AI literacy with STEM subjects, environmental education, or local cultural studies. In traditional schooling contexts, the use of AI tools has demonstrated significant potential in STEM education. These tools have been shown to improve learning achievement, motivation, and higher-order thinking skills like problem-solving and computational thinking, through personalized learning environments, intelligent tutoring systems, and adaptive robotic tutors [95].