Higher Education and the Age of AI Linguistic Affordances of Students using ChatGPT in Pakistan

Artificial intelligence tools like ChatGPT have paved the way in Pakistan's higher education. This shift highlights the need to understand the use of language by students, particularly with the assistance of AI (Jamil, 2021; Tayan, et al., 2024). Although affordance theory explains human–tech...

Täydet tiedot

Bibliografiset tiedot
Päätekijä: Anwar, Tahira
Muut tekijät: Humanistis-yhteiskuntatieteellinen tiedekunta, Faculty of Humanities and Social Sciences, Kieli- ja viestintätieteiden laitos, Department of Language and Communication Studies, Jyväskylän yliopisto, University of Jyväskylä
Aineistotyyppi: Pro gradu
Kieli:eng
Julkaistu: 2025
Aiheet:
Linkit: https://jyx.jyu.fi/handle/123456789/103683
Kuvaus
Yhteenveto:Artificial intelligence tools like ChatGPT have paved the way in Pakistan's higher education. This shift highlights the need to understand the use of language by students, particularly with the assistance of AI (Jamil, 2021; Tayan, et al., 2024). Although affordance theory explains human–technology interaction well (Gibson, 1979; Norman, 1988; Gaver, 1991), there are limited studies testing how someone’s disciplinary background affects the linguistic aspects of interaction outside the Western context. This study attempts to fill the gap by addressing: (1) What key linguistic affordances do students of social, and engineering sciences demonstrate during their interactions with ChatGPT in educational settings? And (2) how are these affordances similar and different among students of social sciences and engineering sciences? The study employed a mixed-methods content analysis (Krippendorff, 2018; Saldaña, 2021), 620 original student-given prompts to ChatGPT (n=100 per discipline). A five-category coding scheme, grounded by affordance theory, along with research on chatbots (Kakhki et al., 2024; Jeon, 2024), was used. Results reveal that students from both disciplines primarily use question-based and command-based prompts, while giving tasks mainly related to brainstorming, writing assistance, explanation, and prompts were typically moderate in difficulty, often containing hedging, technical jargon and several steps of scaffolding (Zhong et al., 2023). Social sciences students preferred reflective and open-ended prompts, whereas those in engineering sciences wanted topical questions where the evidence was clear. These findings show that prompt formulation is a linguistic way to demonstrate perceived AI affordance and is mainly influenced by discipline norms. Higher education institutes should enhance AI literacy by designing programs that equip students to use AI tools productively. Moreover, policy reforms are needed to integrate AI in the education system constructively, rather than seeing it as a threat to academic integrity.