A conceptual exploration of generative AI-induced cognitive dissonance and its emergence in university-level academic writing
Top Cited Papers
Open Access
- 17 June 2025
- journal article
- research article
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 8, 1573368
Abstract
University-level academic writing is a form of scholarly communication that demands precision, clarity, and adherence to established conventions. It is a skill fundamental to science education, best honed through continuous practice (Moskovitz & Kellogg, 2011). However, rising academic demands and evolving educational environments, such as shifts to online learning, introduce new challenges. Students often view writing as a daunting task, largely due to insufficient instruction bridging technical composition and creative expression in their coursework (Stride, 2024). Non-native speakers find this struggle compounded, grappling with the linguistic precision and stylistic conventions of academic writing (Nazaroff, 2011). Furthermore, traditional academic writing has also suffered from an overemphasis on technical correctness at the expense of fostering creative expression. In many cases, the perceived marginal role of writing in research and teaching has led educators to delegate writing instruction solely to language departments (Alley, 2024). Such delegation has resulted in a fragmented approach to teaching writing skills, leaving many students underprepared. In the advent of Generative Artificial Intelligence (GenAI), some students now assume that technology can fully compensate for their underdeveloped writing abilities, inadvertently undervaluing the importance of building a strong writing foundation.GenAI tools-such as ChatGPT™ by OpenAI, Gemini™ by Google, and Claude™ by Anthropichave emerged as practical aids for organizing and simplifying writing tasks such as article reviews, lab reports, and research papers (Essel et al., 2024). These technologies help reduce the cognitive load by shifting focus from the mechanical aspects of writing to higher-level critical analysis and interpretation (Olatunbosun & Nwankwo, 2024). However, their integration into academic practice often creates an ambiguous boundary between human intellectual effort and machine assistance (Amoozadeh et al., 2024), hence leading to what we term as cognitive dissonance (CD), a classic social psychology theory introduced by Festinger in 1957. CD is the inconsistency in thoughts, actions, or behavior that leads to a tension or discomfort from holding contradictory beliefs (Hilberg, 2017). This tension is resolved by changing our thoughts or behaviors, adding a new thought, or rationalizing the inconsistencies (Oxoby & Smith, 2014).To date, no consensus has been reached regarding the manifestation of cognitive dissonance in the integration of GenAI in university-level academic writing. Here, we provide a novel perspective on how CD emerges with the integration of GenAI in university-level academic writing. We posit that GenAI both triggers and exacerbates pre-existing tensions, reshaping academic writing dynamics. Furthermore, exploring pedagogical frameworks like constructivism may offer pathways to mitigate these emerging challenges and help students integrate GenAI tools responsibly.Cognitive dissonance has long been a topic of concern in academic writing. Traditionally, it refers to the psychological tension experienced when students' ideals clash with the practical constraints they face (McGrath, 2017). Operationalized as a state of psychological discomfort (Vaidis & Bran, 2019), CD in academic writing often emerges from the conflicting demands between producing high-quality, original work and managing limited time, resources, or guidance. The magnitude of the dissonance experienced depends on the importance of the conflicting cognitions and the proportion of dissonant versus consonant thoughts related to a specific behavior or belief (Festinger, 1957). When faced with this discomfort, individuals are strongly motivated to reduce it. Classic dissonance reduction strategies include: (1) changing one of the dissonant cognitions (e.g., altering one's attitude towards academic integrity), (2) changing behavior to align with cognitions (e.g., stopping the use of GenAI in ways perceived as dishonest), ( 3) adding new consonant cognitions to justify the behavior (e.g., emphasizing the time saved allows focus on higher-level thinking), or (4) trivializing the conflict itself (e.g., deciding that the specific assignment is not important anyway) (McGrath, 2017;Stephens, 2017). Understanding these mechanisms is crucial for analyzing how students might react to the internal conflicts provoked by GenAI use in academic writing. For example, ethical dilemmas may arise when the pressure to publish meets the rigorous standards of originality and research integrity (Schrems & Upham, 2020). This internal conflict can lead students to adopt selective coping strategies, such as focusing only on information that confirms their existing beliefs (Metzger et al., 2020), or in some cases, resorting to dishonest practices (Stephens, 2017). In traditional academic settings, CD can serve as both a source of psychological stress and a catalyst for personal growth. This occurrence prompts students to recalibrate their beliefs and strive for higher standards. However, the introduction of GenAI into this landscape introduces new dimensions of dissonance that are still not fully understood.A trigger, in this context, refers to any factor that disrupts an individual's core beliefs and values, creating psychological tension. GenAI's efficiency, while beneficial, directly conflicts with the academic values of originality, effort, and intellectual ownership. Empirical studies (Chan, 2024;Playfoot et al., 2024) provide evidence that GenAI use can lead to what some describe as "AIgiarism" -a nuanced form of plagiarism where students, while disapproving of overt AI-generated content, find themselves ambivalent about employing AI for paraphrasing or idea generation. For instance, Chan (2024) found significant ethical ambivalence regarding AI use for idea generation versus direct submission. Moreover, Playfoot et al. (2024)...Keywords
All Related Versions
This publication has 24 references indexed in Scilit:
- Students’ perceptions of ‘AI-giarism’: investigating changes in understandings of academic misconductEducation and Information Technologies, 2024
- Generative AI in higher education: Seeing ChatGPT through universities' policies, resources, and guidelinesComputers and Education: Artificial Intelligence, 2024
- Experimental evidence for the efficacy of generative AI in improving students’ writing skillsQuality Assurance in Education, 2024
- Exploring students’ perspectives on Generative AI-assisted academic writingEducation and Information Technologies, 2024
- Hey ChatGPT, give me a title for a paper about degree apathy and student use of AI for assignment writingThe Internet and Higher Education, 2024
- ChatGPT effects on cognitive skills of undergraduate students: Receiving instant responses from AI-based conversational large language models (LLMs)Computers and Education: Artificial Intelligence, 2023
- Dealing with dissonance: A review of cognitive dissonance reductionSocial and Personality Psychology Compass, 2017
- How to Cheat and Not Feel Guilty: Cognitive Dissonance and its Amelioration in the Domain of Academic DishonestyTheory Into Practice, 2017
- Cognitive Dissonance and “The Will to Believe”Fudan Journal of the Humanities and Social Sciences, 2016
- Why we writeIndoor Air, 2011