An article on recognising texts generated by artificial intelligence has been published in ‘The Economist’. Dr Karolina Rudnicka from the Faculty of Languages at the University of Gdańsk provided expert commentary.

In the article, which explores the challenges associated with recognising texts created by large language models (LLMs), dr Rudnicka points out that there is no single, universal ‘AI style’. Just as people differ in their writing styles, individual language models can also exhibit distinctive stylistic features.

‘There is no single AI writing style, just as there is no single human writing style,’ emphasises dr Karolina Rudnicka in a statement quoted in the article.

Dr Karolina Rudnicka’s research focuses, amongst other things, on linguistic variation, the impact of new technologies on communication, and the ways in which artificial intelligence models use language.

The article ‘How to spot AI writing?’ highlights that, as technology advances, the line between text created by humans and text generated by artificial intelligence is becoming increasingly difficult to discern. It adds that analysing individual words and phrases is not sufficient to unequivocally determine the authorship of a text. Instead, context, familiarity with the author’s style, and the way in which artificial intelligence tools are used are becoming increasingly important.