DW ReCo has been involved in several projects that address the question of how to best support journalists, fact checkers, and other stakeholders in their verification workflows. news-polygraph is the latest R&D effort in this context. It aims to combine artificial and human intelligence in the process of content analysis.
How does it work?
The consortium wants to build a multimodal, interoperable platform driven by the latest in AI tech. For this purpose, various tools and methods for content analysis (text, audio, image, video, multimodal) as well as extended services will be combined in a modular system architecture. Human intelligence will complement the platform. To be more specific, journalists will be able to involve third parties with different areas of expertise. Those "crowd workers" will assist with a specific task, e.g. the geolocation of a picture.
This journalist / AI / human expert approach is (currently) a unique characteristic of news-polygraph. The platform has the potential to speed up, simplify and template the verification process and become a versatile tool that can support many different use cases. In the wake of advanced (generative) AI systems used to churn out mis- and disinformation, a project like news-polygraph is highly needed in the verification community.
What's DW's contribution?
DW Innovation will support news-polygraph in several ways: We focus on the needs of investigative journalists, who may face terabytes of data that require special attention to privacy and security. We'll also experiment with deepfakes, explore how the crowd can support investigative research, and coordinate user requirements.
Who's funding this?
news-polygraph was selected for funding by the Bundesministerium für Bildung und Forschung (BMBF, i.e. the Federal Ministry of Education and Research) as part of the RUBIN program (Regionale Unternehmerische Bündnisse für Innovation i.e.: Regional Entrepreneurial Alliances for Innovation").