The linguists and computer scientists of the University of Szeged are working on the development of a humane and automatic fake news recognition system – the public relations directorate of the higher education institution informed MTI.
According to the announcement, researchers from the Department of General Linguistics of the University of Szeged Faculty of Humanities and Social Sciences have been investigating the problem of manipulation in various forms of language use for years.
Quoting academician T. Enikő Németh, they wrote: the spreaders of fake news use powerful manipulation to persuade readers to buy certain products or become supporters of certain conspiracy theories. These texts are characterized by features of language and language usage that articles or news articles describing real scientific results do not have. For example, fake news is characterized by the excessive use of imperatives, verbs of fact, or the emphasis on fear-inducing expressions – the linguistics professor listed.
The Hungarian Academy of Sciences (MTA) has launched three major national research programs. One of these is the National Program Science for the Hungarian Language, which includes a sub-program dealing with the linguistic identification of fake news and pseudo-scientific views.
The consortium implementing the project is managed by the University of Szeged, and linguistic researchers and computer scientists from the universities of Szeged and Debrecen participate in the work. The experts focus on fake news spreading on the Internet, including those related to health care, and in addition to the vocabulary of the texts, they also examine the grammatical and pragmatic tools.
In addition to the scientific theoretical foundation of the concept of fake news, their goal is to develop a human fake news recognition procedure. In addition, automatic tools are being developed, such as a mobile phone application and a browser extension, which will detect the elements that characterize fake news within the checked texts. The manipulative product description revealed by linguistic methods is followed by a professional health inspection. This results in a convincing procedure, a possibility of practical use, the application of which can even save lives – emphasized the academician.
Starting from the theoretical foundations of science, through application to innovation, the program “Linguistic identification of fake news, pseudo-scientific views” extends to the support of which the MTA allocates a total of HUF 200 million. The academy’s national language research program started on December 1 last year and will end on November 30, 2026, the announcement states.