Universities globally continue to escalate their fight against academic misconduct with AI detection tools, integrity departments and updated misconduct policies. But evidence from 3,070 students who gave their reflections in an academic integrity module suggests we’re addressing the wrong problems.
An Australian university conducted an analysis of mandatory reflections from students who were found to have committed some form of academic misconduct between 2019 and 2021.
This analysis reviewed approximately 1.5 million words from students after they faced disciplinary consequences ranging from warnings (40% of cases) through mid-level sanctions to severe outcomes, including subject failure.
These 500-word reflections, required as part of the post misconduct educational intervention, were reviewed by staff to ensure authenticity and offer important insights into how students understood their own misconduct after being caught breaching academic integrity policy.
The data collection ended in 2022, just before generative AI made such authentic reflections difficult to verify.
The timing was fortuitous as these student voices were captured in the final moment before AI transformed the academic misconduct detection landscape forever.
Three uncomfortable truths
Three persistent themes in the analysis emerged across both pre-pandemic and COVID-era reflections, suggesting enduring challenges rather than temporary disruptions.
• Truth #1: They genuinely didn’t understand the rules: ‘I did not know’ dominated the reflections – not as excuse, but as a genuine lack of understanding.
Students described being overwhelmed by policy documents during orientation, misunderstanding collaboration boundaries and discovering only after sanctions that practices accepted elsewhere constituted misconduct at this university.
They understood blatant plagiarism was wrong. But sharing notes? Recycling their own work? Paraphrasing without proper attribution? The reflections revealed students learning these rules only after breaking them.
One pattern emerged repeatedly: students who thought they were demonstrating good teamwork were actually colluding.
This confusion transcends borders.
Research from the United Kingdom, Canada and Australia consistently shows students from different educational cultures struggle to navigate Western academic integrity concepts – particularly around collaboration, citation and intellectual property.
These students weren’t claiming ignorance to escape responsibility – they were expressing genuine surprise at the breadth of what constitutes academic misconduct.
• Truth #2: ‘No Time’ meant crisis, not procrastination: Time pressure emerged as the second dominant theme, but not in the expected way.
These weren’t procrastinators facing predictable consequences. The reflections described students who started assignments early, but faced impossible convergences: employers threatening termination, family medical emergencies, multiple major assessments due simultaneously.
Students viewed these time pressures as a cause of their misconduct. They described being caught between failing to submit (guaranteed failure) and submitting work that wasn’t entirely their own (possible detection). Under pressure, many gambled on the latter.
This aligns with global research linking employment obligations and academic workload to misconduct. But the reflections add nuance: these students weren’t managing time poorly; they were managing impossible situations poorly.
When adverse life events collided with inflexible deadlines, academic integrity became collateral damage.
• Truth #3: They believed they couldn’t succeed legitimately: The most troubling theme was self-efficacy – students’ fundamental belief about their capability to complete assessments successfully.
The reflections revealed students convinced they were inadequate, too embarrassed to seek help or certain everyone else understood what they didn’t.
Students described being paralysed by anxiety, unable to ask questions that might expose their confusion. This existed before COVID-19, but it intensified during lockdowns when isolation amplified insecurity.
They chose risking misconduct charges over perceived humiliation.
The reflections connected this directly to mental health struggles. Students described anxiety, depression and stress as factors in their decision-making. When students believe they cannot succeed through legitimate means – whether this is accurate or not – cheating becomes survival strategy rather than moral choice.
What makes these findings credible?
Crucially, these reflections showed little evidence of neutralisation – the rationalisation typically found in cheating research.
Few students refused responsibility or minimised their actions. Perhaps because sanctions had already been imposed, defensive postures had given way to genuine reflection.
Students expressed recognition of wrongdoing and eagerness to avoid future misconduct.
This authenticity was verified through institutional oversight. Reflections suspected to be non-genuine resulted in students repeating the post-misconduct educational intervention, ensuring the dataset represented authentic student voice rather than strategic responses.
These findings suggest three fundamental shifts:
• From information dump to just-in-time education: Deliver integrity education when students need it – before first essays, at group project starts, when learning discipline-specific conventions. Make it mandatory but brief, practical not theoretical.
• From rigid systems to structural flexibility: Acknowledge students have complex lives beyond university. Create legitimate extension pathways. Stagger assessment deadlines. Recognise that inflexible deadlines disadvantage students managing employment, family responsibilities and health challenges.
• From assuming competence to building confidence: Create low-stakes opportunities for practice and safe failure. This assists with the transition to any university. Normalise help-seeking through peer mentoring and proactive support. Reach struggling students before desperation drives cheating.
These 3,070 students have given higher education an invaluable gift: honest insight into why academic misconduct occurs from those who engaged in it and experienced the disciplinary consequences. Their reflections suggest misconduct often stems from confusion, crisis and lack of confidence rather than laziness or moral failure.
As generative AI revolutionises both cheating and detection, these pre-AI student voices become even more valuable.
They reveal that the roots of academic misconduct run deeper than technology, and touch upon how and when we educate about integrity, how we structure assessments and how we support students who are struggling.
Universities worldwide face a choice: continue escalating the technological arms race or address the fundamental issues these reflections revealed. Because if students don’t understand rules, can’t manage competing pressures and don’t believe they can succeed, no amount of investment will preserve academic integrity.
Dr Albert Munoz serves as student ombudsman at the University of Wollongong, Australia. Dr Mercedez Hinchcliff serves as associate dean of education in the faculty of business and law at the University of Wollongong. Dr Jonathon Mackay is a senior lecturer at the school of business, University of Wollongong. Professor Ann Rogerson serves as chair of the Academic Integrity Advisory Group at the University of Wollongong.
