Table 5 Qualitative themes, participant... | Emerald Publishing

Table 5

Qualitative themes, participant coverage and representative quotations (n = 8)

Theme	Analytic meaning	Participant coverage and coded references	Main pattern and variation	Representative quotation
AI as an academic support tool	Generative AI was used to generate ideas, locate references, prepare presentation outlines and accelerate routine academic work	Seven-eighths participants; 17 coded references	AI use was most readily accepted for preparatory and lower-stakes tasks. It was positioned as a support tool rather than as the final producer of academic work	“AI helps in searching for initial research ideas.” (P02)
Task-selective adoption	Acceptance depended on the type of task and the level of academic responsibility involved	8/8 participants; 21 coded references	AI was more acceptable for presentations, reference searching and teaching materials. Journal articles, research and academic books required stronger verification, rewriting and human control	“For presentations, it is acceptable, but journal articles still need verification.” (P01)
Ethical ambivalence and epistemic risk	Perceived efficiency coexisted with concerns about plagiarism, dependency, declining critical reading, changes in information-seeking behavior and blurred authorship	6/8 participants; 16 coded references	Some participants emphasized risks to authorship, independent reading and critical information-seeking, while others viewed AI as a neutral technology whose consequences depended on how it was used	“AI can blur the boundary between academic assistance and plagiarism.” (P03); “AI itself is not the problem.” (P02); “Some students use AI before reading sources, changing how they seek and assess information.” (P06)
Critical verification and professional responsibility	Lecturers maintained academic judgment through fact-checking, source validation, rewriting and alignment with their own arguments	6/8 participants; 13 coded references	AI was accepted as process support, but lecturers retained responsibility for validity, originality and the final academic argument	“For journal articles, everything still needs to be verified, rewritten and adjusted according to our own academic arguments.” (P01)
Institutional governance gap	The absence or ambiguity of institutional policy required lecturers to establish their own boundaries for acceptable AI use	Five-eighths participants; 11 coded references	Participants from public and private institutions described unclear regulations, although the form of institutional support could differ	“Our university still does not have clear regulations regarding AI use.” (P04)
Limited formal training support	Formal AI training opportunities remained limited and were not always aligned with lecturers' professional needs	4/8 participants; 8 coded references	The interviews described limited access to formal training; differences in training needs by seniority were established primarily through the survey	“AI training opportunities are still very limited.” (P07)

Theme	Analytic meaning	Participant coverage and coded references	Main pattern and variation	Representative quotation
AI as an academic support tool	Generative AI was used to generate ideas, locate references, prepare presentation outlines and accelerate routine academic work	Seven-eighths participants; 17 coded references	AI use was most readily accepted for preparatory and lower-stakes tasks. It was positioned as a support tool rather than as the final producer of academic work	“AI helps in searching for initial research ideas.” (P02)
Task-selective adoption	Acceptance depended on the type of task and the level of academic responsibility involved	8/8 participants; 21 coded references	AI was more acceptable for presentations, reference searching and teaching materials. Journal articles, research and academic books required stronger verification, rewriting and human control	“For presentations, it is acceptable, but journal articles still need verification.” (P01)
Ethical ambivalence and epistemic risk	Perceived efficiency coexisted with concerns about plagiarism, dependency, declining critical reading, changes in information-seeking behavior and blurred authorship	6/8 participants; 16 coded references	Some participants emphasized risks to authorship, independent reading and critical information-seeking, while others viewed AI as a neutral technology whose consequences depended on how it was used	“AI can blur the boundary between academic assistance and plagiarism.” (P03); “AI itself is not the problem.” (P02); “Some students use AI before reading sources, changing how they seek and assess information.” (P06)
Critical verification and professional responsibility	Lecturers maintained academic judgment through fact-checking, source validation, rewriting and alignment with their own arguments	6/8 participants; 13 coded references	AI was accepted as process support, but lecturers retained responsibility for validity, originality and the final academic argument	“For journal articles, everything still needs to be verified, rewritten and adjusted according to our own academic arguments.” (P01)
Institutional governance gap	The absence or ambiguity of institutional policy required lecturers to establish their own boundaries for acceptable AI use	Five-eighths participants; 11 coded references	Participants from public and private institutions described unclear regulations, although the form of institutional support could differ	“Our university still does not have clear regulations regarding AI use.” (P04)
Limited formal training support	Formal AI training opportunities remained limited and were not always aligned with lecturers' professional needs	4/8 participants; 8 coded references	The interviews described limited access to formal training; differences in training needs by seniority were established primarily through the survey	“AI training opportunities are still very limited.” (P07)