RuralWell AI: A Hybrid AI and Physical System for Supporting Well-Being in Rural Students

UNAM (Universidad Nacional Autonoma de Mexico)

Haziel, Álvarez González, UNAM; Luis Torres Lozano, UNAM; Hans Zúñiga García, UNAM; Norma Elva, Chavez, UNAM; Saiph, Savage, UNAM and Northeastern University

Daily life / student life / campus

We are designing a low-cost support system for students in the Global South, with an initial focus on rural regions in Mexico. The goal is to help identify when students may be experiencing stress and to provide forms of support that can help them better manage it. The system has two main parts: an online AI support system and a low-cost autonomous physical system. The online AI system will act as a daily companion for students. It will interact with them through lightweight check-ins, short conversations, and simple support around schoolwork and daily challenges. As part of these interactions, the system will deliver surveys through an online agent. These surveys will be informed by existing literature on student stress, allowing the system to more systematically understand students’ stress levels. Based on students’ responses and patterns of interaction, the system will provide appropriate recommendations and activities to help them manage stress. These may include breathing exercises, reflection prompts, study organization strategies, journaling activities, and other low-pressure practices that support both well-being and learning. When needed, the online system will also help connect students with volunteer psychologists through social media channels that are already familiar and accessible to them. This creates a simple layered model of support in which the AI system helps identify possible stress and offer immediate guidance, while human volunteers can provide additional support when necessary. In parallel, we will create a low-cost autonomous physical system that provides tangible, offline resources for rural students without requiring internet access. This physical system can be placed in schools or community spaces and serve as a local hub where students can borrow and return materials that support emotional well-being and stress management. The physical system could include books, emotional well-being workbooks, games, creative materials, and stress-reduction tools that students can use on their own or with peers. For example, it could include storybooks and culturally relevant reading materials that support reflection and resilience, guided activity booklets on mindfulness and stress management, board games and puzzles that encourage relaxation and social connection, and notebooks or drawing materials that help students express emotions. It could also include calming materials such as coloring sheets, tactile objects, and reflection cards. This approach matters because it combines online and offline support in a way that fits the realities of rural communities. Many systems assume stable internet access, but students in underserved regions may need support that is available both digitally and physically. By combining an online AI system with an autonomous physical resource hub, we create a practical and accessible model that can continue supporting students across different levels of connectivity. Together, these two components create a clear and grounded support system: the online AI system helps assess stress and guide students toward helpful activities and human support, while the physical system ensures that students also have offline resources they can access in their everyday environments.

Rural students in Mexico, and across the Global South more broadly, often face high levels of stress shaped by poverty, long travel times to school, limited educational resources, family responsibilities, and uncertainty about future opportunities. Their stress is not only academic. It is often tied to everyday material and social pressures. At the same time, they usually do not have the same support systems available to students in cities. Many rural communities have limited access to counselors, psychologists, extracurricular programs, reliable internet, and digital mental health tools. This means students may experience significant stress but have few resources to help them manage it. Their stress also differs from that of students in the Global North and urban settings. In those contexts, stress is often framed around grades, competition, and social pressure. For rural students in Mexico and the Global South, stress is more often shaped by structural inequality, economic hardship, weak infrastructure, and limited institutional support. This use case matters because it focuses on students who face serious stress but are often overlooked by existing support systems. It calls for tools designed around their realities, not around the assumptions of urban or high-income contexts.

Our motivation is to design AI systems that meaningfully support students who are often overlooked by existing technologies, particularly rural students in Mexico and across the Global South. Much of today’s AI for education and well-being is built around assumptions of stable infrastructure, abundant resources, and access to professional support. We are motivated to challenge these assumptions by creating systems that reflect the real conditions students face, including limited connectivity, fewer institutional supports, and different forms of stress. We are also driven by a commitment to human-centered and worker- and community-centered AI. This project extends that approach to students, focusing not only on learning outcomes but also on well-being, emotional support, and equitable access to resources. Our learning objectives are threefold. First, we aim to understand how stress is experienced and expressed by rural students, and how it can be responsibly assessed through lightweight AI interactions and survey-based methods grounded in existing research. Second, we seek to explore how AI systems can provide meaningful, context-aware support without over-reliance on constant connectivity or high-cost infrastructure. Third, we want to study how combining online AI systems with low-cost physical resource hubs can create more resilient and inclusive support models. Through this work, we hope to generate both practical system designs and broader insights into how AI can be used to support well-being in underserved communities in ways that are culturally grounded, accessible, and sustainable.

Situation or context We address the lack of accessible, context-appropriate support for managing stress among rural students in Mexico and, more broadly, across the Global South. These students often face high levels of stress driven by economic precarity, long travel distances to school, limited educational resources, and family or work responsibilities. At the same time, they have limited access to mental health services, school counselors, or reliable digital tools that could support their well-being. Critical analysis Existing AI and digital well-being tools are largely designed for urban and Global North contexts. They often assume stable internet access, individual device ownership, and the availability of professional support systems. As a result, they do not translate well to rural settings, where connectivity is inconsistent and support infrastructures are minimal. Additionally, many systems treat stress as an individual and short-term issue, rather than as something shaped by structural conditions such as inequality and limited opportunity. This creates a mismatch between the design of current tools and the lived realities of rural students. Team perspectives and debates Within our team, we discussed different approaches to addressing this gap. One perspective emphasized building a fully online AI system that could scale easily and provide personalized support. Another perspective raised concerns about over-reliance on connectivity and highlighted the importance of designing for offline access and shared community resources. We also debated how to balance automated support with human involvement, particularly in sensitive areas such as mental health. Through these discussions, we converged on a hybrid approach that combines AI-based support with pathways to human assistance and offline resources. This approach reflects a shared agreement that effective systems must be both technologically scalable and grounded in local realities. Proposed contribution and conditions for implementation Our contribution is a hybrid support system composed of two integrated components. First, an online AI system that engages students through conversations and survey-based check-ins grounded in stress research, enabling the system to assess stress levels and provide tailored recommendations. Second, a low-cost autonomous physical system deployed in schools or community spaces that provides offline resources such as books, games, and creative materials that support stress management. This system could be implemented in collaboration with local schools, community organizations, and volunteer networks of psychologists. Key conditions for implementation include access to basic digital devices for students when online interactions are needed, partnerships to ensure culturally relevant content and responsible use of stress assessment methods, and local support for maintaining and distributing physical resources. Under these conditions, the system can provide a practical, scalable, and context-aware way to support student well-being in underserved communities.