Winter School on Artificial Intelligence for Language Learning, Teaching, and Assessment

Key points emerging from the different sessions:
An anthropocentric view of AI
A recurring theme, introduced by coordinator Letizia Cinganotto, is that "there is nothing artificial about AI": it is created, developed, used, and governed by humans. This "humanistic" approach suggests that AI should be a tool to enhance human abilities, not replace them, always keeping the individual at the center of the decision-making and creative process.
Language assessment and professional standards
Nick Savil (ALTE/Cambridge) illustrated the evolution of language assessment in the digital age:
Assessment cycle: AI is already used to automate processes such as writing evaluation (automated rating) and the generation of test materials.
Learning-oriented assessment (LOA): Technology enables us to overcome the divide between formative and summative assessment, creating "smart" assessment models that are adaptive, diagnostic, and integrated into the learning journey.
Equity and social justice: Savil highlights that access to technology is a matter of social justice. It is essential to ensure that AI does not exacerbate the digital divide and that it also supports less widely spoken languages.
European policies and multilingualism
Anna Soler (European Commission) presented the EU policy context:
1+2 Objective: The EU aims for every young person to know at least two languages in addition to their school language, but data show that there is still a long way to go, especially for languages other than English.
AI as support: AI is seen as a potential tool for the personalization of teaching and as support for linguistic mediation in multilingual classrooms, for example helping immigrant students to integrate.
Research projects and practical applications
The seminar is part of the IMPACT project (Mediterranean partnership for collaborative teaching), which aims to promote internationalization and the use of advanced digital infrastructures. Some practical examples discussed include:
Chatbot ID: A virtual assistant developed for conversational practice in Italian.
Computational lexicography: The use of AI to create Arabic-Italian dictionaries from parallel corpora, facilitating the learning of a complex lexicon.
Padlet: Used during the Winter School as a central hub for sharing resources, assignments, and case studies among participants.
Luciano Floridi’s philosophical perspective
Professor Floridi proposed a vision of education as "linguistic competence":
Semantic capital: Education serves to cultivate an individual's semantic capital, that is, the body of knowledge and narratives that give meaning to life.
KFL (Knowledge as a Foreign Language): Floridi suggests treating knowledge as a foreign language: the more "languages of information" (from mathematics to biology) we speak, the more conscious citizens we become.
Distant Writing: Floridi imagines a future in which AI will write most of our texts (emails, reports), while humans will focus on the design, control, and responsibility of content, acting as "architects" of information.
Ethical challenges and regulation
Giorgio Resta analyzed the European Union's AI Act, highlighting that AI systems used in education are often classified as "high risk." This entails strict obligations in terms of transparency, human oversight, and prevention of bias in data.