Unlocking the potential of AI and EdTech in Mexican Schools

The education system in Mexico faces significant challenges, including educational inequality, inaccessibility, skills qualification disparities, and the lack of standardized tools to measure student progress and detect learning disabilities early. In response, our project explores the transformative potential of AI-powered EdTech to enhance educational outcomes and provide early mental health indicators, particularly in resource-limited settings. Conducted in collaboration with Amy Orben’s group at Cambridge, our research comprises two main components.

The first component, developed in collaboration with Gillian R. Hayes and Stephen Schueller from UC Irvine, focuses on Amira Learning—an AI tutoring platform designed to improve reading skills in public schools. Deployment of Amira Learning yielded significant gains in reading fluency and comprehension, demonstrating its efficacy as a personalized learning tool. However, its impact on reducing reading-related anxiety was modest, indicating that while AI tutors are effective for skill acquisition, they should be integrated with additional socio-emotional support measures.

The second component involved the development of a real-time Android app called iTACO, created in collaboration with Cambridge. iTACO is designed to distinguish between habitual and non-habitual phone users by capturing mobile-touch interactions immediately after unlocking the phone—recording gestures such as taps, swipes, scrolls, and drags. In the first year, after securing ethical approval from both CICESE and the University of Cambridge, we recruited 30 adolescents aged 18–22 to install iTACO on their phones for a week, with all data anonymized and securely hosted on a CICESE server. Further studies with over 100 participants in both Mexico and Cambridge revealed that frequent app switching and erratic tap interactions are correlated with increased anxiety levels, suggesting that digital behavior patterns may serve as passive markers for early detection of mental health issues.

Overall, our integrated research underscores the potential of AI and EdTech in bridging educational disparities and enhancing early mental health screening. While Amira Learning demonstrates promising results in literacy enhancement, the additional behavioral insights provided by iTACO offer a novel avenue for understanding and addressing anxiety. For broad deployment of such technologies, it is crucial to invest in robust technological infrastructure, comprehensive teacher training, and stringent ethical guidelines to safeguard data privacy. These findings pave the way for innovative, data-driven educational interventions that not only improve academic outcomes but also support holistic student well-being.

Open-source protocol:
Feel and Touch: A mobile game to identify digital markers of ASD.
iTACO: Interactive Touch-Based Application for Conductual Observation

Publications

Valdez C., Alvarez-Molina, K., Castro, L. A., & Tentori, M. (2023). The Doppler Effect: Altering Pitch to Materialize Movement Direction in Yoga. Conference on Human Factors in Computing Systems (CHI’23).

Téllez, A. M., Castro, L. A., & Tentori, M. (2023). Developing and Evaluating a virtual reality videogame using biofeedback for stress management in sports. Interacting with Computers, iwad025.

Arturo Morales Tellez, María Concepción Valdez Gastelum, Luis A. Castro, and Monica Tentori, 2023. Evaluating the Effect of the Color-Word Stroop Test and VR as a Psychological Stressor. In Adjunct Proceedings of the 2023 ACM International Joint Conference on Pervasive and Ubiquitous Computing.

Luis M. Zamudio, Ivonne Monarca, Jesus Favela, and Monica Tentori (2023). Toward using Digital Phenotypes in Mobile Tactile Stimulation for Children with ASD. In Adjunct Proceedings of the 2023 ACM International Joint Conference on Pervasive and Ubiquitous Computing the 2023 ACM International Symposium on Wearable Computing (UbiComp/ISWC ’23 Adjunct ).

Monarca I, Cibrian Franceli L., Tentori Monica. Can knowledge representation support the design of AI systems for children with autism spectrum disorder? CHI 2023 Workshop on Child-centred AI Design: Definition, Operation and Consideration

Ivonne Monarca, Edgar Chavez, Franceli L. Cibrian, Monica Tentori. Using a Small Dataset to Classify Strength-Interactions with an Elastic Display: A Case Study for the Screening of Autism Spectrum Disorder (Research highlight). Poster presentation . CHI 2023 Workgroup on Interactive Systems in Healthcare

Grants

Digital Habits: Phenotyping Mobile Touch Interactions to Identify Habitual Phone Use in Adolescents (CICESE – University of Cambridge). Submitted to the Economic and Social Research Council (ESRC)

Exploring the effect of Amira in Reducing Reading Anxiety and Improving Reading Ability in Mexican Children, Accepted as CERES Subgrant (CICESE-UC Irvine)