Philologists Created an Automatic Service to Replace "Toxic" Expressions

09:57, June 10, 2024

It avoids the risks of conflicting communication

With the development of digital technologies, people have become sensitive to the incorrect use of words, which has given rise to many conflicts in the media and social networks. To mitigate this trend, specialists from Ural Federal University developed a new service that identifies potentially "toxic" expressions and replaces them with more acceptable ones. Anna Plotnikova, the head of the study and Director of the Faculty of Philology at UrFU, spoke about the new service at a press conference at TASS. The service was developed with financial support from the Priority-2030 program.

Now the service is being tested with an IT company. Once it has been fine-tuned, the developers plan to make it available to the public. Philologists hope that the application will help to make communication more correct, since under the influence of digitalization in the media and social networks the rules of communication and speech etiquette have begun to change.

"We begin to use shorter messages, and we are also not always greeting or saying goodbye to the addressee as we used to. As a result, a special form of communication appears - oral-written. Formally it is written, but in some respects, it is oral, because when communicating in messengers, we do not always put punctuation marks, write with mistakes, or use constructions that are characteristic of oral speech, but we use them in the written form of communication," explains Anna Plotnikova.

The emergence of a large number of neural networks capable of generating text has led to the widespread use of so-called synthetic or artificial texts. Many people, especially those involved in advertising, create written content that has no authorship in the classical sense. The expert notes that the spread of this trend is a serious challenge for the whole society.

"Such texts can be called pseudo-human, as they are created by artificial intelligence. In this regard, we can say that there is a problem with the authorship of such content. Such texts have no authorship, but at the same time, they are based on large language models that are created and trained by humans. And the problem of such artificial, synthetic texts will grow, especially in advertising and marketing," concludes Anna Plotnikova.

Sergey Lukyanchenko

The news article was translated by UrFU International Marketing Office.
info.international@urfu.ru, +7 (343) 375-45-19