Detailed description of project

AI4Debunk aims to develop innovative AI-powered tools to support trustworthy online activity among citizens and effectively debunk disinformation. The project focuses on creating four human-centered AI-powered interfaces: a web plug-in, a collaborative platform, a smartphone app, and an augmented reality (AR) interface. These tools will be built upon a robust “debunking” API designed to identify and challenge false information.

To develop the debunking API, AI4Debunk will focus on two primary case studies: the disinformation surrounding the war in Ukraine and climate change. By conducting a deep analysis of these topics with the help of sociological experts, the project will map disinformation content onto knowledge graphs. These graphs will be compared to existing databases using AI methods such as natural language processing (NLP). The resulting knowledge graphs will serve as the foundation for technical developments, enabling the creation of multimodal and monomodal debunking modules capable of handling various types of content, including written text, images, and videos.

The debunking modules will use advanced AI and machine learning methods, including recurrent neural networks (RNN), convolutional neural networks (CNN), and transformer-based models, to detect and analyze questionable content. These modules will generate a “Disinofscore” to rate the fakeness of the content. The API will integrate this scoring mechanism, providing the backbone for the four interfaces. The web plug-in will alert users to fake or true content in real-time, the collaborative platform (Disinfopedia) will allow proactive users to report questionable content for verification by a debunking committee, the smartphone app will enable users to check content on the go, and the AR interface will offer insights into emerging fake news trends on social media.

Type and scope of work provided

Project Management: AI4Debunk will provide comprehensive project management services, ensuring effective coordination among all consortium members, adherence to project timelines, and the achievement of project goals. This includes regular meetings, progress tracking, and reporting to the funding authority.

Communication, Dissemination, and Engagement: The project will implement a robust communication strategy to disseminate information and engage with various stakeholders. This includes creating and distributing content across multiple platforms to raise awareness about the project’s objectives and findings. The dissemination efforts will target the general public, academic communities, and policymakers to ensure widespread impact.

Research and Development: The core service provided by AI4Debunk involves extensive research and development activities. This includes the creation of AI-powered tools and interfaces, the development of the debunking API, and the construction of knowledge graphs. The research will leverage advanced AI techniques and methodologies to ensure the accuracy and reliability of the debunking tools.

Technical Infrastructure and Support: The project will provide the necessary technical infrastructure to support the development and deployment of the AI-powered tools. This includes setting up data processing environments, integrating IoT devices, and ensuring the scalability and security of the developed systems.

Training and Capacity Building: AI4Debunk will offer training programs and workshops to enhance the media literacy and critical thinking skills of various target groups. These programs will educate participants on identifying disinformation and using the project’s tools effectively. Training sessions will be tailored for different user groups, including journalists, educators, and the general public.

Pilot Testing and Validation: The project will conduct pilot testing of the developed tools in selected regions to validate their effectiveness and gather user feedback. This phase will involve real-world testing scenarios to ensure the tools’ usability and impact. Feedback from pilot tests will be used to refine and improve the tools before broader deployment.

Policy Recommendations and Advocacy: Based on the project’s findings, AI4Debunk will develop policy recommendations to support the creation of a safer and more reliable online information environment. These recommendations will be shared with policymakers and other stakeholders to influence policy changes and promote the adoption of best practices in combating disinformation