В УрФУ разработали сервис, помогающий заменять в тексте токсичные слова

The development helps identify "bad" words and replace them with "good" ones

Philologists from Ural Federal University (UrFU) have developed a service to identify and replace toxic words in texts in order to avoid accidental insults. This tool is a plugin for Microsoft Word that allows users to check their texts for words and phrases that could cause negative reactions. The service offers neutral replacement options for words that may have been used unintentionally, which helps improve the quality of communication, reduce conflicts, and increase tolerance.

Professor Anna Plotnikova of the Department of Fundamental and Applied Linguistics at UrFU noted that the service helps authors detect conflictogenic elements in the text. For example, discussing migrants in a politically incorrect way can be problematic, and the service offers to replace such expressions with more neutral ones. Such programs can soften the polemic, preventing lawsuits on issues of insults and protection of honor and dignity.

The service takes into account the context of word usage, which is important for proper understanding and replacement. For example, the words "old men" and "old women" can be offensive if used in relation to elderly people, but will be appropriate in a quote from works such as Pushkin's "The Tale of the Fisherman and the Fish." The program analyzes the immediate environment of words to determine their meaning and suggest correct replacements.

The program does not replace words automatically, but only offers replacement options, leaving the author free to choose. The main goal of the service is to make communication successful, conflict-free, and harmonious.

The service was developed as part of the project "Human Well-being in the Context of Digitalization." This is the first such tool for the Russian language, although there are already analogues on the market for the English language, such as Angry Translator, which translates from "angry" language to "kind" language. The development will soon be available for free use.

Read materials on the topic:

"Yandex Keyboard" learned to correct and improve texts

Will shorten the text and highlight the main idea: VK neural network added to Mail.ru service

YandexGPT neural network learned to summarize text, anyone can test the service

Sources
«RG»

Now on home