Inclusive communication is essential for the development of a democratic society, as it ensures social cohesion and guarantees equal opportunities for all, especially in an increasingly diversified society. However, most languages in the world do not include stereotypes, do not have a gender-neutral form or do not allow to address people with disabilities in an inclusive way.
The use of language is a fundamental tool for communication and the basis for social cohesion. In order to ensure equality of opportunity, it is therefore essential that communication is inclusive, i.e. that it does not exclude any social group. As the world becomes increasingly diversified, it is more important than ever to ensure that communication is inclusive. The term encompasses the inclusion of groups such as minorities, women, and people with disabilities.
The project aims to contribute to the development of NLP-based methods to detect, highlight, and correct stereotypes, gender bias, and ableism in text. Those tools leverage the annotation provided by domain experts (e.g., linguistics and sociolinguistics) to let deep learning models learn to detect and correct those language phenomena. The project has started with a specific focus on the Italian language and will be extended to a larger set of languages (e.g., French, Spanish etc.).
Example Gender neutral form (Italian):
The project requires the participation of linguistics experts and computer scientists, who will work together to develop the required tools. The multidisciplinary nature of the project will provide an excellent opportunity to train future computer scientists in linguistics, and vice versa.
Project team: