Description du projet
Combiner l’intelligence artificielle et humaine pour lutter contre la désinformation en ligne
Les fausses nouvelles, la désinformation et les discours haineux en ligne sont omniprésents et exacerbent les tensions sociétales. Face à la crainte croissante que cette rhétorique néfaste n’incite à la violence et ne pousse à l’extrémisme, une stratégie mondiale pour contrer ce phénomène numérique est plus importante que jamais. Dans ce contexte, le projet HYBRIDS, des Actions Marie Skłodowska-Curie, s’efforce de fournir aux chercheurs les connaissances et les outils dont ils ont besoin pour combattre la désinformation. Plus précisément, il intégrera les connaissances structurées fournies par les sciences sociales et humaines dans des outils de traitement du langage naturel et des algorithmes d’apprentissage profond. En développant de nouveaux systèmes d’intelligence hybride, HYBRIDS combinera l’intelligence de la machine et l’intelligence humaine pour surmonter les lacunes des méthodes d’intelligence artificielle existantes.
Objectif
False rumors, fake news or hate speech against vulnerable minorities through social media are becoming one of the main threats against democracies. A global strategy against disinformation is crucial as open democratic societies rely on free citizens with the ability to access verifiable information so that they can build their own opinions on different political issues. The main scientific objective of HYBRIDS is to provide researchers with the knowledge necessary to design strategies and tools to response to disinformation on the basis of a deep analysis of public discourse. There have been interesting advances in the automatic detection of disinformation by making use of natural language processing and new artificial intelligence techniques in the specific field of machine and deep learning.
However, this is a very complex task that requires a high degree of natural language understanding, inference and reasoning. To improve the strategies to deal with disinformation, HYBRIDS will integrate the structured knowledge provided by social and human sciences into natural language processing tools and deep learning algorithms, so as to develop new hybrid intelligence systems. The concept of Hybrid Intelligence consists of combining machine and human intelligence to overcome the shortcomings of existing artificial intelligence methods.
Even though hybrid systems are likely to become more and more decisive in the immediate future, there are very few experts who can design and develop such systems. This is mainly due to the multidisciplinary nature of the hybrid strategy and the difficulty of finding researchers with full training in traditionally separate disciplines such as computer engineering, social sciences or linguistics. We believe it is time to launch a DN with the necessary requirements to train researchers in hybrid methodologies for their application in social studies with an impact on the sustainability of good democratic practices in Europe.
Champ scientifique
- humanitieslanguages and literaturelinguistics
- engineering and technologyelectrical engineering, electronic engineering, information engineeringinformation engineering
- natural sciencescomputer and information sciencesdata sciencenatural language processing
- social sciencespolitical sciencesgovernment systemsdemocracy
- natural sciencescomputer and information sciencesartificial intelligencemachine learningdeep learning
Mots‑clés
Programme(s)
- HORIZON.1.2 - Marie Skłodowska-Curie Actions (MSCA) Main Programme
Régime de financement
HORIZON-TMA-MSCA-DN - HORIZON TMA MSCA Doctoral NetworksCoordinateur
15782 Santiago De Compostela
Espagne