Between Promise and Peril: Artificial Intelligence as a Contested Catalyst for Research Development in Tunisia

Artificial intelligence (AI) is increasingly positioned as a catalyst for research development in the Global South, yet empirical evidence on how researchers themselves engage with it remains scarce, particularly in Maghrebian contexts where the literature has concentrated on teaching and student support rather than scholarly practice. This study addresses that gap through a mixed-methods design that examined AI adoption, its epistemological implications, and its institutional-ethical conditions among 100 researchers in Tunisian higher education. Data were collected via a structured online questionnaire combining closed-ended items (analysed descriptively using SPSS [Version 27]) and open-ended items (analysed through Braun and Clarke’s six-phase thematic analysis), interpreted through strategic-applicative, philosophical-systemic, and socio-technical-ethical lenses. Results show that AI use is widespread but discipline-shaped: engineering and exact-science researchers favour modelling and simulation, while humanities and social-science researchers favour translation and synthesis. Half of the respondents (50%) identified declining creativity and critical thinking as the principal obstacle to use, positioning epistemic rather than technical concerns at the centre of adoption. Strong normative support for transparency (80% favouring attribution of AI-generated content) coexists with a marked demand for institutional guidance, revealing a behavioural-normative asymmetry between cautious use (65% occasional) and ethical expectation. These findings indicate that AI’s developmental value is conditional: access alone does not translate into epistemic autonomy unless accompanied by discipline-sensitive literacy, contextually grounded tools, and explicit governance. The study contributes an empirically grounded account of AI’s dual promise and peril in a Global South research system, with implications for research policy, AI literacy, and Open and Distance Learning practice.