Rethinking algorithmic violence and agency in the post-humanist era: Grok’s toxic communication crisis

This study examines the July 2025 Grok toxic communication crisis as a case through which algorithmic violence, platform governance, and digital agency became publicly contested. The empirical material consists of 30 anonymized Turkish X posts identified through searches for “#GrokKrizi” and “grok krizi” in X’s Top tab during the 20–25 July 2025 collection window, together with five Grok-generated responses elicited between 28 and 30 July 2025 after thematic coding had been completed. The study first uses Braun and Clarke’s six-phase thematic analysis to identify recurring patterns in the user-post corpus, and then applies Fairclough’s critical discourse analysis to selected extracts and to Grok’s responses as discursive artifacts. Five themes were identified: Insensitivity to Cultural Norms, Biased Discourse Targeting Social Groups, Toxic Language Directed at Political Figures, Algorithmic Insensitivity to Social Traumas, and Algorithmic Agency and Ethical Responsibility. The findings show that the Grok crisis was publicly framed not only as a technical malfunction, but also as a problem of platform amplification, governance failure, cultural violation, and distributed responsibility among developers, platforms, users, and algorithmic systems. The article’s main contribution is to demonstrate how post-humanist theory can be used cautiously, after empirical analysis, to interpret AI systems as contested discursive participants without attributing consciousness, intention, or human-like moral agency to them.Consequently, the study demonstrates that artificial intelligence systems should be regarded as discursive actors bearing ethical and social responsibility within digital communication processes.