Fraunhofer develops software for automatic fake news detection
Fraunhofer FKIE has developed a new machine learning tool that analyzes social media posts to automatically detect fake news and disinformation. The software can also identify hate speech.
Germany – The Fraunhofer Institute for Communication, Information Processing and Ergonomics (FKIE) has developed a software system aimed at combating disinformation spread online. The tool analyzes social media posts, such as tweets, and uses machine learning techniques to automatically identify fake news and disinformation.
The system examines both the content of posts and their metadata, including publication frequency and user networks. The analysis relies on machine learning models trained to recognize linguistic markers like unusual phrasing, grammatical errors, or atypical punctuation. These factors can indicate that a news item has been manipulated or fabricated.
Professor Ulrich Schade and his research group have built the tool using training sets composed of both credible news articles and content identified as false by users. The software is also capable of detecting hate speech, as fake news often incorporates such elements to sow distrust or attack individuals or groups.
The software developed by Fraunhofer FKIE can be customized to meet specific customer needs. It can function as an early warning system for authorities or businesses to counter the spread of fake news and disinformation on social media.