Ai and human created media texts: experiment results

Abstract

Many journalists and experts are concerned about how widespread of AI-machines may affect the world of information. Changes can affect both the journalists themselves and the audiences. There are already a number of projects in Kazakhstan introducing the AI usage, however, they do not relate to journalism. Texts created by AI are still not typical for the domestic media sector. In order to understand how Kazakhstani consumers of information relate to such texts, we conducted an experiment based on an electronic questionnaire (the main research method). There was no data in the domestic scientific literature on the question whether people are able to distinguish the text written by AI from the text written by a real journalist. This research question formed the basis of our experiment. The purpose of the study is to determine the qualities of a journalistic text that make it possible to distinguish it from a robotic text. The scientific significance is due to the relevance of the topic, the in- terdisciplinary direction of work. Practical significance lies in the ability to use data in several scientific fields. The results show that the majority of media consumers (135 participants out of 153) were able to distinguish two texts and identify which of them belongs to the robot. Additionally, according to the an- swers, Kazakhstani mediaconsumers are not ready for the appearance of AI texts in the mass media, they consider them poor quality, incomprehensible, logically unrelated and, in principle, respondents prefer texts written by a person – a professional journalist. Most likely, this trend will continue over the next several years. This work contributes to solving the problem of using AI in the media sphere. Key words: journalism, AI, media text, mass media, robo-journalism.