The spread of disinformation is posing increasingly significant challenges for journalism. In local journalism in particular, there is often a lack of resources to thoroughly verify information and assess its reliability. At the same time, independent and fact-based reporting is a key prerequisite for the functioning of democratic societies.
Against this backdrop, the collaborative project “PROVAIDE” is investigating how journalistic work can be specifically supported through the use of artificial intelligence. The focus is on developing an AI-supported tool that helps journalists to systematically analyse information and better assess its credibility.
The tool analyses the origin, dissemination channels and typical patterns of information and assesses the likelihood of disinformation. The multimodal evaluation of text, image, video and audio data enables a particularly well-founded assessment. In addition, a dynamic fact database is being built up, which can be continuously expanded and improves the quality of the analyses.
There is a particular focus on practical applicability: the solution developed is intended to integrate seamlessly into existing editorial workflows and provide tangible support to journalists in their day-to-day work. ConPolicy brings its expertise in user and behavioural research in the digital world to bear here.
Furthermore, the project takes an innovative approach that goes beyond mere content analysis. It also takes into account the origins and dissemination dynamics of information in order to make the evaluation of sources more transparent and traceable. The open-source provision of key findings, as well as close collaboration with media organisations and research institutions, help to ensure long-term use and further development.
With PROVAIDE, ConPolicy, together with its project partners, is contributing to strengthening quality, transparency and trust in the digital information space.
The project will be implemented across eleven work packages:
- Project management
- Needs analysis
- Crawling, data infrastructure and data preparation
- Extension of crawling to include multimodal data
- Expert fact annotation
- Design of the user interface (front-end)
- Development of AI backend and user frontend, including operation throughout the project duration
- Extension of the demonstrator to include multimodality
- User-centred evaluation of the demonstrator
- Concept for subsequent use and continued operation
- Science communication
Within the project, ConPolicy is responsible for leading the design of the user interface and for the user-centred evaluation of the demonstrator. This encompasses both qualitative and quantitative surveys, as well as the derivation and development of design proposals for the front-end. Furthermore, the design, implementation and analysis of the evaluation are among the key tasks.
