AI is everywhere. But what about integrity?

In late July 2025, the University of Cape Town (UCT) made headlines when it announced its decision to stop using Turnitin’s artificial intelligence (AI) detection tool. The stated reason? The software was producing too many false positives and false negatives, wrongly accusing students of using generative AI (genAI) when they had not, or failing to detect genAI use when it had occurred. While this decision might appear justified given the imperfect nature of the technology, it also signals a deeper unease: the erosion of credible systems to uphold academic integrity in the generative genAI era.

Within the learning environment more shaped by genAI, the role of institutions is to work through the complex exercise of ensuring academic integrity under conditions of flawed detection aids and institutional level unreadiness. This will require a multi-faceted, forward-thinking response in which technological innovation and pedagogical adaptation are weighed against ethical responsibility to facilitate genuine learning and unbiased assessment. At the heart of such a strategy is the establishment of accurate, context-specific standards for the use of genAI, in combination with curricula that promote genAI literacy and critical thinking rather than passive replication of information.

Institutions must also prioritize developing comprehensive genAI policies that directly govern its proper use in academic, administrative, and pedagogical spheres. These laws should address basic concerns such as information privacy, equal access to genAI technology, and algorithmic discrimination, hence safeguarding the rights of the learner and locking in safe and open learning environments. Given the pace of AI advancement, such policies should be dynamic, in continuous revision, and reactive in character to tackle emerging challenges and opportunities.

Moreover, creating genAI literacy on all levels of stakeholders’ students, workers, and the administration is key to empowering ethical and informed engagement with genAI technologies. To this effect, traditional mechanisms of academic honesty must shift from a reactive, detection-based approach to a values-driven culture that advocates for transparency, accountability, and human-centric genAI competencies. This shift places integrity not merely as compliance, but as a shared institutional culture resilient against the chaos engendered by genAI.

Successful response to the revolutionary influence of genAI on scholarly practice needs systems thinking and facilitation from all levels of the scholarly ecosystem: students, faculty, administrators, and IT support. Academic integrity is not merely detection programs or disciplinary policies in the current era; it is cultural transformation toward ethical genAI interactions facilitated by structural, pedagogical, and policy-based innovations.

One of the major drivers to making academic integrity a reality in the age of genAI is restructuring assessment. Traditional assessment that relies heavily on rote memorization, formulaic essays, or easily replicable structures is vulnerable to genAI-generated answers. However, real world assessments that reflect real life tasks, personal reflection, and require connections between ideas within fields are more impervious to genAI abuse. For example, oral defences, project learning, portfolio submissions, and scaffolded research journals can demonstrate deeper understanding and unique student voice.

Teachers must be empowered to move in this direction. GenAI-aware curriculum planning, critical thinking facilitation, and ethical technology adoption professional development courses must be provided. Lecturers and tutors must be empowered to design learning experiences involving genAI constructively teaching students how to utilize tools like ChatGPT for brainstorming, research assistance, and iteration, and establishing boundaries and expectations.

Institutions must build robust genAI governance structures that reflect institutional values, comply with regulation, and adapt in responding to evolving genAI realities. An effective policy framework should contain statements on genAI usage, which anticipate students and staff to clarify whether and how genAI tools were used in creating scholarly work. This responsibility and reflective practice are based on transparency. Moreover, academic misconduct rules need to be stratified with a distinction between exploratory abuse and malicious deception to facilitate instructional interventions on early-stage abuse.

In addition, policy development must consider data privacy, accessibility, and ethical procurement and deployment of genAI resources. Disproportionate access to sophisticated genAI services can serve to widen the current inequalities in tertiary education, and institutions must have systems in place to prevent unfairness and inequity. Above all, policies for genAI must be living documents subject to ongoing review and amendment through feedback by diverse groups of academic stakeholders like students, academic staff, and technologists.

Integration of genAI literacy as a graduate skill is at the heart of future-proofed education. Like digital literacy, information literacy, and business communication skills being what are considered essential, so also must genAI literacy be seen as a competency core. This means understanding the abilities and limitations of genAI, the moral implications of its use, and the broader societal impact of algorithmic decision-making. Institutions can enable this by developing cross-disciplinary modules in genAI, promoting interdepartmental learning, and using peer-led discussion forums that examine ethical issues raised by genAI applications.

Academics must also model ethical genAI use demonstrating when and where genAI software is implemented in research, communication, or curriculum planning. Leadership commitment matters most; if senior leaders visibly support responsible genAI innovation and invest in hardware and professional development, they make genAI literacy and integrity institutional values everyone embraces.

While resources such as Turnitin and GPTZero are constantly evolving, they should not be employed as the primary means of ensuring integrity. Instead, they should be employed formatively improving student learning and understanding of originality. Additionally, collaboration among institutions can provide a shared expertise base of best practice and resource sharing to improve responses to genAI-related challenges.

Finally, the incorporation of generative genAI is not an assault on academic integrity but an opportunity to reenvision it. Educational institutions must meet the challenge with transparency, imagination, and moral determination, developing responsive pedagogies, welcoming policies, and human-oriented academic cultures. Through doing so, they will not only secure the integrity of their qualifications but also equip students to flourish as moral, adaptive professionals in a genAI-advanced world.

25 August 2025

Authors: Dr. Johannes Sekgololo, Tania Homan and Nelson Phiri