1. Survey sample and statistical claims. You state over 500 participants but never report the exact N, response rate, or sampling strategy. The percentages (~70%, ~45%, ~20%, ~60%) are presented without confidence intervals or subgroup breakdowns (students vs. staff). More seriously, these percentages sum to well over 100%, yet you never clarify whether categories are mutually exclusive or overlapping. How can readers assess the reliability of these figures?
2. Table 1 is malformed and unreadable. On page 4, the overview of methodological approaches is rendered as a single run-on line of text, not a table. The columns (Research phase / Methodological approach / Goal) are not separated. This makes the methodological overview impossible to parse and undermines a core claim of transparency.
3. Internal contradiction in the autoethnography section. You write: the authors did not use GenAI to create any sentences or paragraphs for the original, non-edited draft of this paper (p. 6). Yet earlier in the same paragraph you state you used GenAI to proofread parts of the text, create English translations of the German quotes, and extend our ‘manual’ literature review. This is a distinction without a clear difference: proofreading and translation do modify sentences. The claim is semantically evasive and should be stated precisely.
4. Epistemic messiness is not adequately distinguished from existing concepts. You define it as a state of the academic assemblage caused by two seemingly incompatible tendencies… that simultaneously strengthen and weaken epistemic relations (p. 2). This is functionally equivalent to longstanding STS concepts such as sociotechnical ambivalence, interpretive flexibility, or epistemic uncertainty. What does epistemic messiness add that these do not? The paper never clarifies.
5. The assemblage framework is applied but never tested. You claim GenAI causes deterritorialization that may outweigh its ‘territorializing’ forces (p. 14). This is a purely rhetorical assertion. No metric, observation, or comparative baseline is offered for what would count as outweighing. The framework functions as a descriptive vocabulary, not an analytical tool.
6. Ethical oversight of the media diary method is absent. A customized ChatGPT prompted students with questions about their GenAI use. You do not report IRB/ethics approval, informed consent procedures for this AI-mediated data collection, or how you handled the possibility that the AI’s prompts shaped the very experiences being studied. This is a major methodological and ethical gap.
7. No datasets were generated or analysed (Data Availability, p. 16) is factually false. You conducted 36 interviews, 75 media diary entries, and a survey of 500+ participants. This statement contradicts the entire Methods section and raises serious questions about data integrity and journal vetting.