AI-Testing Assessment Pattern for Database Students
A study published on arXiv (2608.12351) discusses a novel assessment framework for a large second-year undergraduate module on database systems, specifically tackling the challenges that generative AI (GenAI) presents to unsupervised online evaluations. Named the X1-X2-X3 pattern, this design mandates that students source an answer, create their own response, and critique the sourced information. During its development, the assessment underwent rigorous testing with current GenAI tools, leading to modifications when generic prompts yielded superficially acceptable responses. The paper utilizes archived assessment materials, rubrics, planning documents, GenAI trial data, practice responses, achievement records, and feedback from external reviews. Its primary contribution is the X1-X2-X3 pattern for evaluating students' AI literacy, learning outcomes, and reflective skills.
Key facts
- Paper on arXiv:2608.12351
- Focuses on assessment design in the GenAI era
- Designed for a large second-year undergraduate database systems module
- Uses X1-X2-X3 assessment pattern
- Pattern includes sourced answer, own answer, and evaluation of sourced output
- Assessment tasks were stress-tested against GenAI tools
- Draws on archived assessment materials, rubrics, planning records, and more
- Main contribution is the X1-X2-X3 pattern for testing AI literacy and learning outcomes
Entities
Institutions
- arXiv