Vice-Chancellor’s Welcome Email Returns 100% AI Prediction
Recently, a controversy arose surrounding the Vice-Chancellor’s welcome email at the University of Otago, which was reportedly analyzed using an AI detection tool known as GPTZero. The email, sent to all students at the beginning of the academic year, had the subject line “Student update: 25 February 2026”. The anonymous student who submitted the tip claimed that the tool returned a 100% AI-generated prediction on the message.
Background of the Incident
The student noticed specific linguistic structures within the email that are often associated with AI-generated content. For instance, phrases like “it’s not X, it’s Y” and the use of em dashes were highlighted as typical of generative AI language models. This particular structure was exemplified by the statement: “The people who live around you aren’t just locals — they’re your neighbours.”
The Functioning of GPTZero
GPTZero operates by analyzing text for two main factors: burstiness and perplexity. Burstiness refers to the variation in sentence structure, while perplexity measures how predictable the text is to a language model. Independent testing has shown that GPTZero has around 90% overall accuracy, although its accuracy decreases with mixed AI and human content.
Official Information Act Request
Upon receiving the AI prediction, the student filed an Official Information Act (OIA) request with the University, which is obligated to provide information within 20 days. The student inquired whether AI was used in drafting the email and requested transcripts of the AI conversation, alongside the University’s position on such AI usage.
University’s Response
According to the OIA response, the message was drafted by a member of the University’s Communications team, who utilized an approved internal AI system, specifically Microsoft Copilot, to suggest improvements. These suggestions were then accepted or rejected by the Communications team member before final approval by the Vice-Chancellor.
AI Governance at the University
The University has an AI Governance Policy that mandates that AI systems must augment human capabilities, ensuring that human judgment and oversight remain central to all communications. Additionally, the Staff Use of AI Systems Policy is currently being finalized and will stipulate that AI use must be disclosed when it significantly contributes to a work product or is required by research ethics approvals.
Stakeholder Expectations
Students, as key stakeholders, are considered in these policies. However, when asked if the use of AI in the welcome message would warrant a future disclosure, the University responded negatively. A spokesperson stated that using AI to refine a personal communication, with full oversight, does not create a reasonable expectation for disclosure.
Public Reaction
The anonymous student expressed dissatisfaction, suggesting that the Vice-Chancellor, who reportedly earns $700,000 annually, should not be outsourcing such an important message to AI. In response, a University spokesperson emphasized that the Vice-Chancellor is actively engaged with students and staff, asserting that the welcome message was not solely written by AI but was enhanced with AI software.
Conclusion
This incident raises significant questions about the use of AI in institutional communications and the expectations of transparency regarding AI involvement. As universities increasingly adopt AI technologies, the balance between efficiency and the necessity for human engagement remains a pivotal topic for discussion.
