Credit: AIT / GettyImages
The detection of hate speech is becoming an increasingly topical issue in our society due to its prevalence in social networks and messenger apps. As the communicative exchange between people takes place more often and more broadly in social networks, people often meet with different opinions and points of view. This can lead to fellow human beings being defamed, insulted or even threatened. This so-called hate speech is amplified by filter bubbles: once users have entered these echo chambers, they are shown more and more such content by the algorithm of the respective social network and by automatically generated messages (bots). This gives the impression that hate messages are very widespread. In addition, the spread of disinformation increasingly leads to heated discussions that can trigger hate speech. Hate speech is mainly defined by insulting and threatening people on the basis of characteristic features - this mainly includes sexism and racism.
Automatic detection of hate speech through artificial intelligence (AI)
With the Center for Digital Safety & Security, the AIT Austrian Institute of Technology has a comprehensive research focus on the design of secure digitalisation. Here, machine learning methods (AI) and computational linguistics are used on the basis of the latest technologies and tools are developed that can support the automated detection of hate speech in the future. An international team of data science experts is researching state-of-the-art methods for identifying hate speech on the internet within the framework of national and European research projects and in close cooperation with actors from science, industry and the public sector in order to be able to effectively counter this escalating problem. In the future, these media forensic tools will support authorities, for example, in the evaluation of very large amounts of data from the internet. Because in their daily work, they are now confronted with a flood of messages with potential and possibly criminally relevant hate speech comments.
Current developments in the field of media forensic text analysis are currently taking place under the leadership of AIT data science expert Mina Schütz, who was awarded the AIT Poster Award for Young Talent by the AIT in cooperation with the tech incubator accent and the venture capital fund tecnet equity for her work in 2021, among other things as part of the research project Defalsif-AI. The project addresses the detection of multi-media disinformation and is funded by the Federal Ministry of Agriculture, Regional Development and Tourism (BMLRT) within the security research programme KIRAS.
[Contact]
Mag. (FH) Michael W. Mürling, Marketing and Communications Manager, Center for Digital Safety & Security, AIT Austrian Institute of Technology; E-Mail: michael.muerling(at)ait.ac.at Web: https://www.ait.ac.at/dss/