The future of learning is increasingly defined by artificial intelligence (AI), with one such exemplar being ChatGPT. Although AI as a presence in learning spaces has become widely recognised, its potential as an examination invigilator is the subject of speculation and ongoing research. Might ChatGPT, or comparable AI software, become responsible for overseeing examinations? Outside the limelight of extant debates, let us explore some lesser-accessed angles on this developing phenomenon.
Beyond Surveillance: AI as a Learning Companion in Exams
At first impression, AI as an invigilator may appear restricted to surveillance alone—monitoring for cheating, tracking eye movement, or marking suspicious behaviour. But a closer examination reveals that ChatGPT could revolutionise the fundamental nature of exams. Rather than simply monitoring students, AI could act as a “learning companion” during the exam process. Imagine an AI that not only supervises but also engages with students in real-time, providing hints or clarification during the exam to help students better express their understanding, all while maintaining academic integrity. This could drastically shift the focus of exams from memorisation and time-based pressures to true learning and comprehension.
AI’s Role in Personalised Assessment: Redefining Fairness
A less traditional concept lies in AI’s ability to generate tailored tests that adjust in real-time to the skill set and pace of a student. ChatGPT might, in theory, produce personalised questions based on past responses or areas where the student struggles. This contrasts sharply with static, formal exams often not designed to accommodate varying learning styles or measure individuals accurately. In this system, AI would not just act as an invigilator but also as a curator of the exam experience, ensuring every student is assessed fairly according to their learning path, not a common standard.
The Unspoken Risks of AI Surveillance: The Issue of Data Privacy
Though the prospect of AI surveillance in exams may seem desirable, a largely unspoken concern remains: data privacy. The volume of information AI engines like ChatGPT capture extends beyond mere monitoring for cheating. This information may encompass behavioural trends, thinking processes, or even emotional states during the exam. In an invigilation system powered by AI, such trends could be used to assess a student’s “character” or stress levels. What happens to the data following the exam? How do we guarantee its security and ensure it is not exploited for unintended purposes such as profiling or predictive analysis? These are pressing concerns that must be addressed in the future of AI-based education.
Bias Mitigation: AI as a Double-Edged Sword
The potential of AI to promote fairness is widely discussed, yet its very design may perpetuate biases. Even sophisticated models such as ChatGPT learn from data that may reflect societal inequalities. For instance, an AI trained largely on data from Western educational environments could inadvertently favour students from those backgrounds while underperforming with students from multicultural or regional contexts. If ChatGPT is deployed as an invigilator, it may lack contextual sensitivity and perpetuate inequalities it was meant to reduce. The concept of AI-powered invigilation could unintentionally widen the educational divide between socio-economic groups unless the technology evolves to become both adaptive and culturally sensitive.
Redefining Integrity: The Issue of Trust in AI-Based Exams
The level of trust students place in AI-powered invigilation is another under-explored dimension. How might students respond to being monitored by an AI like ChatGPT, knowing it may report behaviour or answers based on programmed logic? Unlike humans, AI lacks empathy or context-based judgement. An innocent error or momentary distraction could be misinterpreted as misconduct. This raises ethical concerns regarding fairness and transparency. Could prolonged use of AI in exams erode student trust in the examination system if the role of AI is not clearly regulated and communicated?
The Future Role of Human-AI Collaboration: A Paradigm Shift
The future of AI invigilators like ChatGPT may not lie in replacing human supervision, but rather in a blended approach. AI could handle real-time monitoring and flag potential issues based on pre-configured patterns, while human invigilators provide moral judgement and contextual understanding. This collaboration would reduce the administrative burden on educators, allowing them to focus on interpreting results, offering feedback, and mentoring students. Such a model would combine AI’s efficiency and scalability with the empathy and insight of human interaction.
ChatGPT and the Way Forward for Examination
While the concept of ChatGPT as an invigilator is still far from mainstream adoption, its potential to transform examinations is compelling. By turning assessments into adaptive, personalised experiences, AI could foster deeper learning environments. Yet challenges such as data privacy, bias, and trust remain unresolved. To realise such a future, a transparent, balanced, and collaborative partnership between human educators and AI systems is essential to ensure equity and fairness in education.
