The main goal of FAST LISA is to develop a model for data collection, capacity building, and community engagement. It focuses on training ambassadors for online and offline awareness actions that identify and combat hate speech online. Furthermore, the effect of youth speech on AI-based algorithmic detection of hate speech is explored. For this purpose, a framework for identifying youth language biases in datasets is developed. The findings are incorporated into the development of domain-specific machine learning algorithms for classifying hate speech.

Anyone seeking reliable information in times of crisis is confronted with a flood of information on the Internet and in social media. The targeted dissemination of false reports or fraudulent offers of help further complicates this situation. This challenge has been addressed by selected partners from the Qurator Alliance in the Panqura project. Using AI-based curation technologies, a technology platform was created for more information transparency in times of crisis - and beyond.

As part of the research and development work, manual sub-processes of curation were automated with the help of AI technologies. These include, for example, searching and finding, viewing, selecting, ordering, sorting, summarizing, classifying, linking, or visualizing content and data. Further developed as curation technologies, they support the process of creating and maintaining content while at the same time creating an optimal basis for intelligent services such as semantic search, recommendation, or monitoring systems. The aim of the joint project was to develop a technology platform that supports a wide range of information sources, data formats and application scenarios. This modular platform is used to test and index.mdevaluate the newly developed processes.