AI for Linguistic Diversity

Despite efforts to bring AI technology to languages and societies lacking in resources, extractive and exploitative research and development practices may exacerbate inequalities, producing end results that fail to benefit native speakers. The Workshop on AI for Linguistic Diversity (AI4LD) aims to explore alternative approaches to building language resources and tools that consider linguistic diversity as a primary objective. The workshop brings together researchers with a large variety of linguistic backgrounds who investigate the place of linguistic diversity in the digital world, and in AI technology in particular. 

The workshop is jointly organised by IMT Atlantique, France and the University of Amsterdam, and is inscribed within the LiveLanguage initiative that promotes value-sensitive and collaborative methods for the development of digital language resources.

Topics

The topics covered by the workshop include:

• the construction of lexicons for minority and under-resourced languages;

• the collection of language- and culture-specific concepts;

• language modelling bias within AI technology;

• the universality of multilingual language resources;

• ethical approaches to working with linguistic communities;

• machine translation over culturally specific terms and expressions;

• tools for language resource development.

Organisers

• Gábor Bella, IMT Atlantique;

• Paula Helm, University of Amsterdam.