As Generative AI, digital collusion, and contract cheating continue to compromise online assessments, educators face a critical challenge in verifying authentic student knowledge. While returning to traditional paper-based exams mitigates digital threats, it often reintroduces old vulnerabilities, such as peer-to-peer copying, and creates massive administrative overhead when attempting to create multiple test versions.

This presentation explores a hybrid solution to the modern academic integrity crisis: utilizing parametric generation to automatically create unique paper-based assessments for every single student in a cohort for both take-home and in-class tasks. Crucially, the core question structures remain common and strictly aligned with Unit Learning Outcomes (ULOs). By altering specific variables, datasets, and scenarios at the individual level, the assessment eliminates the possibility of both digital generation and analogue copying without compromising standardisation, while maintaining similarity for marking purposes.

Participants will explore the architecture of this task-generation approach, discussing how to scale individualised assessments without overwhelming teaching staff and what digital tools can be used for that. We will examine the implementation of this system in technical environments, demonstrating how randomized, paper-bound variables force students to rely entirely on their own cognitive processes and problem-solving skills, thereby restoring confidence in the assessment’s integrity.