Table 1 Hypothetical opportunities to use the SCP framework to reduce the opportunity for AI-facilitated academic misconduct before, during, and after AI-facilitated misconduct at the University administrative and course coordinator levels
From: Linking artificial intelligence facilitated academic misconduct to existing prevention frameworks
Time period of intervention | University administrative level | Course coordinator level |
|---|---|---|
Before commencing work | • Compulsory training relating to misconduct and use of AI tools • Clear, public rules/policy/guidelines relating to appropriate use of AI tools and academic integrity | • Ensure novel assessment items are set for each iteration of a course • Provide relevant practice and support to maximise the likelihood of student success • Give a range of options for assessment topics to reduce frustration with ‘uninteresting’ assignment topics |
During period for completing assessment | • Mandate use of certain types of technology that monitors student assessment activity (i.e., copy-pasting and general writing behaviour) and authenticity (i.e., plagiarism detection) • Require authenticity declarations on submission of assessment | • Operationalise a whistle-blower capacity within the course to enable anonymous tip-offs about cheating • Require students to undertake and submit formative assessment (early drafts) for assignments • Advertise to students that there will be random viva defences, focused on high performers, high within-unit difference scores on supervised and unsupervised assessment items, and/or unusual metric data (e.g., copy-pasting, general writing behaviour, etc.) |
After submitting assessment | • Enforce and publicise prevention/detection successes • Retain student work for future AI-scanning • Mandate difference score monitoring for each student across supervised/unsupervised assessments through all courses they take • Keep administrative level student specific records about misconduct | • Instigate random viva defences, focused on high performers, high within-unit difference scores on supervised and unsupervised assessment items, and/or unusual metric data (e.g., copy-pasting, general writing behaviour, etc.) • Report misconduct when detected/suspected, making sure there is an administrative trail for individual students |