Offline-First Pwa with Controlled Generative AI Support for Teaching Informatics in a Near-Frontline School

dc.contributor.authorTurchyn Dmytro
dc.contributor.authorТурчин Дмитро
dc.date.accessioned2026-09-23T10:45:03Z
dc.date.available2026-09-23T10:45:03Z
dc.date.issued2026
dc.description.abstractFormulation of the problem. Near-frontline schooling faces a practical barrier that becomes methodological: learning continuity and assessment fairness are undermined when electricity and internet access are unstable, while generative AI can further weaken the validity of product-only grading in programming tasks. This study examines whether an offline-first learning design can preserve core learning actions and provide assessable evidence of work under disruption without normalizing unverified reliance on GenAI.Materials and methods. The intervention was an offline-first Progressive Web App (“Edu Survival Kit”) with local persistence and deferred synchronization, complemented by controlled GenAI tutoring and a process-oriented assessment approach based on artifact traces (versions, short reflections, and verification tests). The pilot involved three school cohorts (Grades 7, 10, and 11; 84 learners received access). Evidence included descriptive platform indicators, an anonymous learner survey (n = 33), teacher feedback (n = 2), and external expert appraisal (n = 1). Data were summarized descriptively and thematically. Results. The offline-first design sustained learning during planned outage conditions and markedly reduced task access time compared to an LMS route in a micro-comparison (sub-second versus tens of seconds). Learners reported high perceived usability and strong perceived support under unstable connectivity and stress; the GenAI hint layer received the highest usefulness ratings. Teachers and the expert confirmed crisis-fit and innovativeness but emphasized the main risk: over-trust in GenAI, which can mask shallow understanding.Conclusion. Offline-first architecture can function as instructional infrastructure in near-frontline settings when paired with an assessment that foregrounds process evidence rather than final products. GenAI support is pedagogically acceptable only under explicit transparency and verification rules and with data minimization.
dc.identifier.citationTurchyn D. Offline-First Pwa with Controlled Generative AI Support for Teaching Informatics in a Near-Frontline School [Техt] / D. Turchyn // Фізико-математична освіта : науковий журнал / Міністерство освіти і науки України, Сумський державний педагогічний університет імені А. С. Макаренка, Фізико-математичний факультет ; редкол.: О. Семеніхіна, М. Друшляк, Francis Kwadwo Awuah та ін. – Суми : СумДПУ імені А. С. Макаренка, 2026. – Вип. 1 (41). – С. 32–36. – DOI: 10.31110/fmo2026.v41i1-05
dc.identifier.doi10.31110/fmo2026.v41i1-05
dc.identifier.orcid0009-0006-1570-3253
dc.identifier.urihttps://repository.sspu.edu.ua/handle/123456789/18773
dc.language.isoen
dc.subjectoffline-first
dc.subjectProgressive Web App
dc.subjectinformatics education
dc.subjecteducation in emergencies
dc.subjectformative assessment
dc.subjectassessment validity
dc.subjectgenerative AI
dc.subjectnear-frontline schooling
dc.subjectпрогресивний вебзастосунок
dc.subjectінформатика
dc.subjectосвіта в умовах надзвичайних ситуацій
dc.subjectформувальне оцінювання
dc.subjectвалідність оцінювання
dc.subjectгенеративний ШІ
dc.subjectприфронтова школа
dc.titleOffline-First Pwa with Controlled Generative AI Support for Teaching Informatics in a Near-Frontline School
dc.title.alternativeOffline-First Pwa з контрольованою підтримкою генеративного ШІ для навчання інформатики у прифронтовій школі
dc.typeArticle
dc.udc.udc373.5.016:004:004.738.5:004.8
Файли
Контейнер файлів
Зараз показуємо 1 - 1 з 1
Ескіз недоступний
Назва:
Turchyn.pdf
Розмір:
4.21 MB
Формат:
Adobe Portable Document Format
Ліцензійна угода
Зараз показуємо 1 - 1 з 1
Ескіз недоступний
Назва:
license.txt
Розмір:
2.9 KB
Формат:
Item-specific license agreed upon to submission
Опис: