A Paradigm for Democratizing Artificial Intelligence Research

Erwan Moreau (1) , Carl Vogel (2) , Marguerite Barry

1 LIPN - Laboratoire d'Informatique de Paris-Nord
2 School of Computer Science and Statistics [Dublin]

This proposal outlines a plan for bridging the gap between technology experts and society in the domain of Artificial Intelligence (AI). The proposal focuses primarily on Natural Language Processing (NLP) technology, which is a major part of AI and offers the advantage of addressing problems that non-experts can understand. More precisely, the goal is to advance knowledge at the same time as opening new communication channels between experts and society, in a way which promotes non-expert participation in the conception of NLP technology. Such interactions can happen in the context of open-source development of languages resources, i.e. software tools and datasets; existing usages in various communities show how projects which are open to everyone can greatly benefit from the free participation of enthusiastic contributors (participation is not at all limited to software development). Because NLP research is mostly experimental and relies heavily on software tools and language datasets, this project proposes to interconnect the societal issues related to AI with the NLP research resources issue.

Domaines: Traitement du texte et du document; Intelligence artificielle [cs.AI]

Erwan Moreau, Carl Vogel, Marguerite Barry. A Paradigm for Democratizing Artificial Intelligence Research. Innovations in Big Data Mining and Embedded Knowledge, pp.137-166, 2019, ⟨10.1007/978-3-030-15939-9_8⟩. ⟨hal-02281202⟩