The usability of ChatGPT as a machine translation technology has attracted many scholars to the extent that it is frequently compared with human translation performance. This study attempts to present a comparison between Zeina Idris' Arabic translation as Human Translation (HT), and ChatGPT as Machine Translation (MT) of Elizabeth Gilbert's “Eat, Pray, Love" as a self-help book. The study adopts Boase-Beier's (2014) cognitive stylistics approach to translation, in which she considers, four view-points to attain style in translation. In her framework, she concentrates on the second and third view-points because they concern the translator and his task; her concern “encompasses the source-text author and the target-text reader to the extent that they impact upon this task" (p. 6). Boase-Beier's cognitive stylistic approach explores theories that relate linguistic choices to cognitive structures and processes. Ambiguity, foregrounding, salience, visibility, metaphor, iconicity, mimesis, and diegesis are examined in both human and machine translation to find out which of them achieves the semantic accuracy and the stylistic aspect of the source text. The results indicate the limitations of machine translation to produce a reliable literary translation. Unlike the human translator whose use of the mind style is balanced and reads Gilbert thought and style.