📄 social_science

Rebuilding Academic Trust in the AI Era: An Integrated Education Governance and Institutional Marketing Framework

This study proposes and empirically validates a multidimensional institutional governance framework comprising ethical transparency, algorithmic accountability, digital AI literacy, and adaptive AI governance, demonstrating that robust governance mechanisms significantly enhance institutional trust and perceived ethicality, which in turn drive students' sustainable AI behavioral intentions and strengthen university brand equity in the Generative AI era.

Original authors: Marwa Khairallah

Published 2026-10-01

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Marwa Khairallah

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the modern university, a quiet tension has taken hold. For decades, the relationship between a student and their school has rested on a simple, unspoken agreement: the student does the work, and the school verifies it. But the arrival of powerful new computer programs, capable of writing essays, solving problems, and generating ideas in seconds, has shaken this foundation. When a student submits an assignment, it is now harder for a teacher to know if the work is truly their own or if it was crafted by a machine. This uncertainty has created a strange situation where students are using these tools more than ever, yet they worry that the system is unfair, that their privacy is at risk, and that the value of their degree is slipping away. The core question facing higher education today is no longer just about how to stop academic misconduct, but how to rebuild the confidence that the institution is a fair and honest place in a digital age.

A researcher at the Faculty of Economics and Management of Sfax in Tunisia set out to understand exactly what makes students trust their universities again in this new environment. Instead of asking students if they simply like using a specific app, the study looked at the school itself. The researcher proposed that trust is not built by the technology, but by how the university manages it. To test this, the researcher surveyed 380 university students who were already using these artificial intelligence tools for their studies. The goal was to see which specific actions by the university mattered most to the students' sense of security and their willingness to use the tools responsibly.

The study identified four main pillars that universities must build to restore trust. The first is ethical transparency, which means the school clearly explains where and how it uses these tools, what data it collects, and what the rules are. The second is algorithmic accountability, which ensures that if a computer makes a mistake or an unfair decision, a human is still there to take responsibility and fix it. The third is digital literacy, where the school teaches students how to understand, question, and use these tools wisely. The fourth, and most critical, is adaptive governance. This means the university does not just write a set of rigid rules and forget them; instead, it constantly reviews and updates its policies as the technology changes, involving students and teachers in the process.

When the researcher analyzed the responses, a clear picture emerged. The most powerful factor driving student trust was the university's ability to adapt its governance. When students saw that their school was actively watching the technology, updating its rules, and listening to feedback, their confidence in the institution grew significantly. The next most important factors were the school's willingness to take responsibility for automated decisions and its clarity about the rules. While teaching students how to use the tools was helpful, it was the least powerful driver of trust on its own. The data showed that even a very knowledgeable student would not trust the system if the university itself seemed opaque, rigid, or unwilling to be held accountable.

The study also found that this trust acts as a bridge to behavior. When students trust their university, they are far more likely to believe the school is ethical. This belief in the school's moral character then directly influences how the students act. Students who trust their institution and believe it is ethical are much more likely to intend to use artificial intelligence in a responsible, sustainable way. In fact, the combination of trust and the perception of ethical behavior explained nearly two-thirds of the variation in how students planned to use these tools. This suggests that the path to responsible use is not through fear or strict punishment, but through a relationship built on clear communication and shared responsibility.

The findings challenge the common approach many schools have taken so far, which often relies on banning tools or using software to detect misconduct. The research suggests that these reactive measures are not enough to rebuild confidence. Instead, universities need to function like a living system that evolves with the technology. The study indicates that when a university treats artificial intelligence as a governance challenge rather than just a technical problem, it can turn a crisis of confidence into an opportunity. By being open about their rules, taking ownership of their systems, and constantly improving their approach, schools can create an environment where students feel safe to learn and innovate. Ultimately, the research concludes that the future of higher education depends less on the sophistication of the tools students use and more on the strength of the trust they place in the institutions that guide them.

Technical Summary: Rebuilding Academic Trust in the AI Era

Problem Statement
The rapid integration of Generative Artificial Intelligence (GenAI) in higher education has precipitated an institutional crisis of academic trust, characterized by an "adoption-trust paradox." While students ubiquitously utilize GenAI tools for academic tasks, this usage coexists with deep ethical skepticism regarding integrity, authorship, and institutional legitimacy. Existing literature predominantly relies on individual-level technology acceptance models (e.g., TAM, UTAUT), which fail to account for the socio-ethical and institutional conditions necessary for trust formation. The central problem addressed is that institutional responses have been fragmented, reactive, and often punitive, creating a regulatory vacuum where student adoption outpaces governance. This study posits that restoring trust requires shifting the analytical lens from individual user acceptance to institutional governance, framing academic trust as a strategic brand asset essential for organizational legitimacy.

Methodology
The study employs a quantitative, cross-sectional research design utilizing Partial Least Squares Structural Equation Modeling (PLS-SEM) to empirically test a multidimensional governance framework.

Sample: Data was collected from 380 university students in Tunisia (public and private institutions) who are aged 18+ and have prior experience using GenAI tools. The sample ensures diversity across disciplines (business, engineering, humanities, etc.) and academic levels.
Instrument: A survey instrument measured seven latent constructs using multi-item scales adapted from validated literature. The four independent variables (governance pillars) are Ethical Transparency, Algorithmic Accountability, AI Literacy, and Adaptive AI Governance. The mediators and outcomes are Institutional Trust, Perceived Institutional Ethicality, and Student Behavioral Intention toward sustainable AI use.
Analysis: Data analysis followed a two-stage protocol in SmartPLS 4. First, the reflective measurement model was assessed for reliability (Cronbach's $\alpha$ , $\rho_A$ , Composite Reliability) and validity (Convergent Validity via AVE, Discriminant Validity via HTMT and Fornell-Larcker). Second, the structural model was evaluated for path coefficients ( $\beta$ ), coefficients of determination ( $R^2$ ), effect sizes ( $f^2$ ), and predictive relevance ( $Q^2$ ). Statistical significance was verified using non-parametric bootstrapping with 5,000 resamples. Common Method Bias was ruled out via Harman's single-factor test and full collinearity VIF assessments.

Key Results
The structural model demonstrates robust explanatory power, with the governance pillars explaining 58.4% of the variance in Institutional Trust, and the combined model explaining 61.2% of the variance in Student Behavioral Intention.

Governance Drivers: All four pillars significantly and positively influence Institutional Trust. Adaptive AI Governance emerged as the strongest predictor ( $\beta = 0.297, p < 0.001$ ), followed by Algorithmic Accountability ( $\beta = 0.258, p < 0.001$ ) and Ethical Transparency ( $\beta = 0.231, p < 0.001$ ). AI Literacy had a significant but smaller effect ( $\beta = 0.137, p = 0.006$ ).
Trust and Ethicality: Institutional Trust exhibits an exceptionally strong positive relationship with Perceived Institutional Ethicality ( $\beta = 0.741, p < 0.001$ ).
Behavioral Outcomes: Institutional Trust directly influences Student Behavioral Intention ( $\beta = 0.286, p < 0.001$ ).
Mediation: Perceived Institutional Ethicality acts as a significant partial mediator between Institutional Trust and Behavioral Intention (Indirect $\beta = 0.337, p < 0.001$ ). This indicates that governance structures build trust, which enhances the perception of the institution's moral standing, thereby driving sustainable behavioral intentions.

Significance and Contributions
The paper claims to make both theoretical and practical contributions by reframing the GenAI challenge as a governance issue rather than a purely technical or pedagogical one.

Theoretical Shift: The study departs from individualist adoption models, demonstrating that trust is an institutional outcome produced by the quality of governance (transparency, accountability, adaptability) rather than a byproduct of user familiarity. It empirically validates that structural safeguards are more critical drivers of trust than individual technical competence alone.
Empirical Validation: It provides robust quantitative evidence for a four-pillar governance framework, addressing a gap in literature that often discusses these concepts normatively without empirical integration.
Practical Implications: The findings suggest that universities must move beyond punitive surveillance and detection mechanisms. Instead, institutions should invest in a "trust infrastructure" characterized by adaptive policy review, clear ethical disclosure, and accountable oversight. For university marketers and administrators, the study argues that "ethical AI" positioning must be backed by auditable governance signals (e.g., appeal pathways, syllabus-level transparency) to enhance brand equity and student loyalty.
Pedagogical Shift: The results support a transition toward "AI-resilient" assessment designs and the integration of AI literacy into curricula, emphasizing critical evaluation and ethical use over prohibition.

The study concludes that rebuilding academic trust is a deliberate process of institutional alignment, where legitimacy is secured through visible, fair, and adaptable governance rather than static prohibition or technical fixes.

More like this