Abstract
Artificial Intelligence is rapidly transforming language learning. The emergence of conversational systems based on Large Language Models (LLMs) such as ChatGPT has created unprecedented opportunities for practising foreign languages through natural interaction. At the same time, important questions remain concerning the educational value, inclusivity, and responsible integration of these technologies into university teaching. UNITE – Universally Inclusive Technologies to Practice English aimed to investigate how AI-powered dialogue systems can support the learning of English as a Foreign Language (EFL) in Italian universities. The project brought together researchers from the Universities of Bologna, Macerata, and Napoli “L’Orientale” with the aim of producing evidence-based knowledge and practical resources for teachers, students, and researchers. A distinctive feature of the project was its focus on inclusivity, drawing on the principles of Universal Design for Learning (UDL) and involving students with disabilities and Specific Learning Disorders throughout the research process.
Results achieved
During the project, more than 400 university students participated in classroom-based experimental sessions involving conversational AI systems, generating one of the first systematically collected corpora of interactions between language learners and AI chatbots. The resulting open-access dataset contains over 300 annotated learner–AI conversations and has been released according to FAIR principles, making it available for future research through international research infrastructures. The project combined learner questionnaires, corpus linguistics, discourse analysis, and AI-assisted annotation to investigate both learners’ experiences and the characteristics of AI-generated language. The findings show that students generally perceived AI chatbots as engaging, accessible, and useful for practising English in a low-pressure environment, while also highlighting important limitations, including overly standardised conversational behaviour and implicit normative assumptions embedded in AI-generated discourse. These results underline the importance of promoting critical AI literacy alongside the pedagogical use of conversational systems. UNITE also produced several methodological innovations. These include an original framework (DIS-TAG) for analysing normative discourse in AI-generated language, an innovative workflow combining Large Language Models with expert human revision for learner-error annotation, and evaluation guidelines for assessing dialogue systems from an inclusivity perspective. Building on these findings, the project developed a comprehensive set of open-access guidelines for integrating conversational AI into university English language teaching. The guidelines provide practical recommendations, classroom activities, prompting strategies, and examples based on authentic learner–AI interactions, supporting teachers and students in using AI critically, effectively, and inclusively. Dissemination and impact Project results have been disseminated through numerous scientific publications, conference presentations, workshops, and public engagement initiatives addressing researchers, teachers, students, and the wider public. The project website (https://site.unibo.it/unite/en) serves as a central hub providing access to publications, datasets, guidelines, teaching materials, and other resources. The methodologies and resources developed within UNITE are already supporting follow-up national and international research initiatives and contributing to ongoing discussions on the role of Artificial Intelligence in language education. UNITE was funded by the European Union – NextGenerationEU under the PRIN 2022 programme of the Italian Ministry of University and Research.Project details
Unibo Team Leader: Adriano Ferraresi
Unibo involved Department/s:
Dipartimento di Interpretazione e Traduzione
Coordinator:
ALMA MATER STUDIORUM - Università di Bologna(Italy)
Total Eu Contribution: Euro (EUR) 210.581,00
Total Unibo Contribution: Euro (EUR) 85.385,00
Project Duration in months: 24
Start Date:
05/10/2023
End Date:
28/02/2026
