Means of communicating human and AI-generated subjectivity in Telegram-discourse

The emergence of AI-generated texts in the media space raises the question of how they differ from those created by humans. Through the linguistic category of subjectivity, a human speaker imbues a text with a sense of personhood, which prompts to investigate how AI imitates this subjectivity. The purpose of this study is to identify a set of subjectivity markers that differentiate between human and AI-generated texts within the genre of a Telegram-post. The human and AI-generated posts are analyzed in Russian. The scientific novelty of the research lies in the categorization of subjectivity markers in texts generated by DeepSeek in Telegram discourse, which contributes to the theory of digital subjectivity. The article determines for the first time that AI-generated texts exhibit an inflated concentration of formal subjectivity markers (deixis, rhetorical devices, emotions) alongside a simultaneous deficit of markers associated with credibility (citation, verifiable data, narratives from personal experience). Furthermore, the study is the first to identify the specific strategy that DeepSeek employs to simulate subjectivity, which manifests in the excessive use of spatio-temporal references, directive constructions, and rhetorical questions to compensate for the absence of an original author’s stance.