HCAI4GEND - Human-Centered Artificial Intelligence for Gender-Sensitive Linguistic Simplification in German, English, and Italian Institutional Communication

HCAI4GEND - Human-Centered Artificial Intelligence for Gender-Sensitive Linguistic Simplification in German, English, and Italian Institutional Communication (PRIN 2022, Prot. 2022N3EH4K) is a collection of multilingual resources developed to support the analysis and production of gender-sensitive institutional communication in Italian, German, and English. These resources provide the linguistic foundation for an open- source, multilingual, human-centered artificial intelligence (HCAI) text editor. The tool identifies potentially gender-discriminatory language in university administrative texts and suggests context-appropriate alternatives to promote more inclusive communication. The collection was developed within the framework of research on institutional language, drawing on Vellutino’s (2018) description of institutional Italian. The project builds on this research through a comparative analysis of university communication in Italy, Germany, Switzerland, and the United Kingdom. The initial phase examined Italian, European, and national legislation on transparency and access to public information from a legal-linguistic perspective. It also analysed university guidelines on gender-sensitive language published in Italian, German, and English. Based on this research, a comparable multilingual corpus of authentic university administrative documents was compiled, including regulations, statutes, calls for applications, forms, and contracts available on university websites. The texts were manually annotated in Doccano using project-specific schemes, with particular attention to morphological, lexical, and structural differences among Italian, German, and English. The resulting resources were designed to document, analyse, and model two main strategies for gender-sensitive institutional communication: Communicating gender equality between women and men through linguistic forms that make the feminine grammatical gender visible; Communicating inclusion through linguistic strategies that neutralise or avoid grammatical gender. The annotation process enabled a systematic comparative analysis of gender-sensitive language across the three languages. HCAI4GEND collection includes six main resources: HCAI4GEND_RAW_CORPORA — a comparable collection of authentic texts from the domain of university administrative communication, including legal and administrative documents characterized by the use of the generic masculine. HCAI4GEND_ANNOTATED_CORPORA — a collection of texts selected from HCAI4GEND_RAW_CORPORA and manually annotated at the morphological and lexical levels according to language-specific and cross-linguistic annotation schemes. HCAI4GEND_GENDER_SENSITIVE_CORPORA — a parallel collection of institutional texts rewritten for gender-sensitive communication on the basis of the texts included in HCAI4GEND_ANNOTATED_CORPORA. The rewritings implement strategies of feminine-gender visibility and gender neutralization. HCAI4GEND_METADATA_CORPORA — a structured descriptive record of the university administrative documents included in the collection, providing information about their institutional, legal, communicative and textual characteristics. HCAI4GEND_ANNOTATION_GUIDELINES — a repository of annotation criteria and linguistic recommendations developed for the identification, classification and rewriting of gender-related linguistic phenomena according to the communication strategies investigated by the project. HCAI4GEND_LEXICON_NOMINA_AGENTIS — a multilingual lexical resource containing institutional and professional role names relevant to university administrative communication, together with the gender-sensitive forms and alternatives considered in the project. References Vellutino, Daniela (2018). L’italiano istituzionale per la comunicazione pubblica. Bologna: il Mulino. Vellutino, Daniela (2024): Quali varietà d’italiano insegnare all’IA? Sviluppo di un modello linguistico per un’intelligenza artificiale Human Centred (HCAI). In: Studi Italiani di Linguistica Teorica e Applicata (SILTA), LIII (3), pp. 558–573. Crestani, Valentina, Pedrini, Giulia, Pellegrino, Ramona, Saracco, Caterina (2026): Gender- inclusive language in university communication: a contrastive analysis of gender-inclusive guidelines in German, Italian, and English. In: Linguistik Online 144(3), pp. 45-58. DOI: https://doi.org/10.13092/fh0p6j53 Vellutino, Daniela, Ingenito, Mafalda, Vitiello, Giuliana (2026): Enhancing University Communication: An Artificial Intelligence-Based Tool for Gender-Inclusive Language. In: Linguistik Online 144(3), pp. 179-189. DOI: https://doi.org/10.13092/p6gc7g68 Vellutino, Daniela, Ingenito, Mafalda, Vitiello, Giuliana (2026): Human-Centered Artificial Intelligence for Gender-Inclusive Language Evolution. In: Linguistik Online 144(3), pp. 163-177. DOI: https://doi.org/10.13092/dzn9bq12 Vellutino, Daniela, Ingenito, Mafalda, Vitiello, Giuliana (2026): Studio linguistico e informatico per lo sviluppo di un’intelligenza artificiale per la comunicazione istituzionale sensibile al genere. In: DALLA NON DISCRIMINAZIONE ALLE PARI OPPORTUNITÀ. Le “sfide” poste dalla transizione digitale ed ecologica Milano Ledizioni LediPublishing Pag.49-62 ISBN:9791256007813; 9791256007820; 9791256007837 Vellutino, Daniela, Ingenito, Mafalda, Vitiello, Giuliana (2026): AI per la comunicazione istituzionale sensibile al genere. In Miriam Voghera; Sergio Lubello (a cura di) Parole per le istituzioni. Riflessioni e pratiche per il rispetto del genere. Pag.103-118 San Cesario di Lecce Pensa Editore. ISBN:978-88-6152-414-9 Vellutino, Daniela (2026) Studio per LINFA, un editor di testo Ia per la comunicazione sensibile al genere In Marina Pietrangelo e Francesco Romano (a cura di) Linguaggio giuridico e inclusione. Costruire cittadinanza in una società che cambia. Pag.107-133 Torino G. GIAPPICHELLI EDITORE. ISBN:979-12-211-6845-7; 979-12-211-6846-4