Andorra proposes criminalising digital violence against women and unveils AI to flag non‑inclusive media language

The Andorran Institute for Women (IAD) will ask the General Council to classify digital violence against women as a criminal offence, citing recommendations from international bodies including the United Nations and the European Union. Judith Pallarés, the institute’s secretary general, said the proposal responds to hostile comments and messages on news sites and social networks that disproportionately target women and “generate a great deal of hate,” and stressed that “digital violence against women is also violence” that requires clear legal regulation.

At the same time the IAD unveiled an internal artificial‑intelligence tool, developed with technology consultancy Dualimind, to analyse inclusive language in Andorran media. The system, intended for the institute’s internal use and not currently available to journalists, scans articles to identify gendered and non‑inclusive language, classifies types of bias and proposes alternative phrasings and recommendations to improve gender representation.

In a pilot run between 12 May and 30 November 2025 the tool analysed 18,930 articles from about ten national publications and detected roughly 37,700 expressions linked to non‑inclusive language — an average of nearly two flagged expressions per article. The most frequent finding was the use of the generic masculine, which accounted for about 85% of detections. References to roles and professions expressed only in the masculine made up roughly 12.6% of cases. Less frequent categories included differentiated treatment, derogatory expressions, conditional female denominations, stereotypes and isolated instances of racism, xenophobia or homophobia.

The IAD says the tool offers concrete alternatives — suggesting collective terms (for example, “student body” or “teaching staff”) and gender‑neutral formulations (“people”, “the public” or “the citizenry”) — to help media produce more inclusive copy. Pallarés emphasised that the institute’s aim is not to police newsrooms but to provide resources, training and constructive options to improve editorial practice and the quality of public communication.

The current iteration does not analyse reader comments or social‑media interactions; expanding the system to include comments is planned for a later phase. Pallarés warned that including comments would likely raise the number of detected racist or xenophobic instances and said the tool will be refined to improve accuracy and usefulness for monitoring and informing policy proposals.

The IAD intends to continue developing the platform and may, in a future phase, consider offering editorial guidance or sharing alternatives directly with media outlets. For now, the institute presents the system as a diagnostic and educational instrument to objectify recurring language patterns and support its statutory role of promoting communications free of bias and stereotypes.

Some commentators and editorials have questioned the feasibility and implications of changing entrenched linguistic norms such as the generic masculine, and raised concerns about potential effects on freedom of expression. Opinion pieces argue that the generic masculine is normative in Catalan and other Romance languages and that efforts to alter established usage risk social and linguistic friction. The IAD says its approach is focused on promoting equality and providing practical solutions rather than imposing sanctions on journalistic content.

By combining a legal proposal on digital violence with the language‑analysis pilot, the IAD says it aims to tackle both the online abuse that disproportionately affects women and the everyday linguistic practices that shape public perception. The institute frames these measures as part of a broader effort to strengthen protection against new forms of aggression and discrimination in virtual environments and to improve how public communication represents all citizens.