In recent years, growing societal attention to gender bias and inequality has prompted Universities to reconsider their institutional practices, including the use of inclusive language in formal communication. As central actors in the production and dissemination of knowledge, academic institutions bear a cultural responsibility to ensure that their communication practices reflect principles of equity and inclusion. This study explores the extent to which written communications addressed to students, either by university faculty or generated by artificial intelligence (AI), adhere to inclusive language practices.
The research adopts a qualitative design based on content analysis and evaluates 63 announcements authored by faculty members from two Italian universities, alongside 39 messages produced by three generative AI systems—ChatGPT Basic, Deepseek, and Gemini 2.0 Flash. The analysis is guided by 16 indicators of inclusive language use, drawn from the guidelines developed by Sabatini (1987) and revised by Somma and Maestri (2020). Each message is classified as inclusive, non-inclusive, or partially inclusive, based on the presence or absence of these indicators.
Findings show that inclusive communication practices are strongly gendered: female faculty members display a much higher rate of inclusive language use (48%) compared to their male colleagues (18%). Notably, no male full professor in the sample used inclusive language. Disciplinary differences also emerge, suggesting that the adoption of inclusive language is influenced by specific cultural norms within academic fields. Among AI systems, Gemini 2.0 Flash produced the highest share of inclusive outputs (38%), followed by ChatGPT (23%) and Deepseek (8%). However, the overall low performance of AI models indicates that inclusive principles are not yet systematically embedded in their training.
This paper contributes to the literature by combining human and AI communication analysis under a unified framework and offering empirical insights into how inclusive language is—or is not—implemented in practice. The results underscore the need for targeted interventions, both educational and technological, to promote inclusive communication in academia. By advancing this dual focus on human and AI-generated texts, the study highlights inclusivity not only as an ethical imperative, but also as a strategic priority for fostering institutional credibility and evolution.