Crowdee, a high-tech spin-off of TU Berlin, contributes its experience and cutting-edge assets in the crowdsourcing domain to national and international research and industry projects. The Crowdee platform enables crowd-based dataset generation, annotation for AI-model training and evaluation, accompanied by customization services, to adjust the Crowdee platform to client-specific requirements for data collection and system integration. Crowdee helps to solve data and AI centic tasks, which deal with the collection and annotation of training and test data and participates in user assessments. In the news-polygraph project consortium Crowdee will leverage their more than 10 years of experience in crowd- and expert-sourcing, including crowd-based innovation and ideation applications. Crowdee has already gained experience and currently takes part in multiple national and EU wide innovation projects.

The German Research Centre for Artificial Intelligence GmbH (DFKI) is Germany’s leading business-oriented AI research organisation. Amongst 28 research departments, nine competence centres and eight living labs, the SLT (Speech and Language Technology) addresses challenges associated with fact-checking, content verification and manipulation detection, and fake news detection across various formats, e.g. audio, image, video, and text. The SLT contributions extend into multilingual speech recognition and speaker verification, integrating crowd-based as well as expert-based AI and human-in-the-loop, focussing on explainability in AI systems, trust, transparency, fairness, and robustness, eventually creating the forefront of innovation in combating misinformation and enhancing media integrity for streamlined journalistic processes.

DW is Germany’s international broadcaster. Our job is to provide unbiased news and information in 32 languages so that people around the world can make up their own minds.

Since the late 1990s, DW Innovation has participated in about 50 nationally and internationally funded R&D projects – many of them in the area of disinformation, such as content verification of text, audio, video and images, as well as bias and deep-fake detection.

As one of the user partners of news-polygraph, DW Innovation focuses on the coordination of user requirements and the evaluation of R&D results. In particular, we are addressing the needs of investigative journalists who may be faced with gigabytes of text, audio, image and/or video data that needs to be verified. Further research focuses on deepfakes and how the crowd can support investigative research.

The Fraunhofer Institute for Digital Media Technology IDMT is one of 76 institutes and research facilities of the Fraunhofer-Gesellschaft, the world‘s leading organization for application-oriented research.

At Fraunhofer IDMT‘s headquarters in Ilmenau (Thuringia, Germany), the experts work on the secure and efficient AI-based recognition and classification of audio and video data for the fields of industrial production, media and environmental monitoring. Another focus is the development of audio technologies for virtual acoustic product experience as well as customized solutions for the production and reproduction of authentic and spatial sound experiences for the areas of professional audio, entertainment, and automotive.

In news-polygraph, Fraunhofer IDMT contributes its expertise and technologies for audio forensics and audio analysis, provenance analysis, visual analysis, recommendation systems and trustworthy AI, and coordinates the R&D activities within the project.

Rundfunk Berlin-Brandenburg (rbb) is the public broadcaster for Berlin and Brandenburg. It offers people in the broadcasting area a comprehensive and balanced range of information, education, culture and entertainment, based on its statutory program mandate. Among other things, rbb uses its regional television program, six radio stations, apps, social media channels, the website rbb24.de, smart speaker services and the rbb Mediathek. rbb is also part of ARD, to whose supra-regional offerings it also contributes programs.
rbb focuses on innovation and innovation in order to continuously meet the challenges of digitalization. For two decades, the broadcaster has successfully acquired research and development projects through its Technical Innovation Management department in order to develop, implement and use the latest media and technology standards for the broadcaster’s programs.
In news-polygraph, rbb is one of the application partners – we identify and define potential areas of application for the emerging platform in order to adapt it to the real needs of future users. In doing so, we pay particular attention to the needs of journalists working on a daily basis. Furthermore, we are involved in the evaluation and piloting of technological developments and are working on the implementation at rbb in coordination with editorial offices and other stakeholders. Our goal here is to develop technical solutions that will reliably support journalists in their work verifying media in the future.

The Quality and Usability Lab is part of the Faculty IV of Electrical Engineering and Computer Science at the Technische Universität Berlin (TU Berlin). The focus of research and teaching at the Quality and Usability Lab is the development of measurement methods for the quality and usability of information and communication technology systems. From this, requirements for system design can be derived, quality and usability can be predicted on the basis of system properties and the methods described can be used in the usability cycle. In addition, methods for natural language processing, anonymization and transparent presentation of AI-generated results are being developed in various third-party funded projects together with different partners.

Ubermetrics is a leading media and data intelligence platform with a focus on research and development. Since 2021, Ubermetrics has been part of the UNICEPTA Group, a leading global provider of media intelligence solutions.

With our specialists, we develop technologies and products to automatically analyze the growing amount of public information from online and offline sources and structure it for the needs of companies. From over 460 million sources, the Ubermetrics platform processes hundreds of pieces of content per second.

Our technologies and innovations have a strong foundation in various disciplines such as natural language processing, data science, information retrieval, information visualization, bioinformatics and UX design. They are developed in close collaboration with internal and external experts in technology, science and business.