Perceived by some people as a threat to academic integrityacademic integrity and ethics and by others as a dependable source of information and inspiration, ChatGPTChatGPT is approached in this chapter from a perspective likely to reveal its real generative capacities in terms of authenticityauthenticity every time a new version is requested, starting with the same prompts within the same session. This evaluation is deemed significantly relevant to the field of text-generating bots because it highlights the obvious limitations regarding text variability, particularly in vocabularyvocabulary choices and grammatical structuresgrammatical structures. The corpus for this study consists of the generation of five sequential texts by ChatGPT within the same session on the prompt ‘define education’education’. The texts are measured for originalityoriginality by using online similaritysimilarity calculators and researcher’s observations. The coefficients used in measuring the originality of each new rewritingrewriting indicate high degrees of similarity between the variants or a total overlapping of the variants, which stands for an exiguous originality. The generative capacity in ChatGPT is a misleading promise of a changed, maybe improved subsequent version when it actually proves a perfect copy with some mere synonym replacement in most cases and few grammatical changes.