Skip to main content

AI coding agents

Permitted assistance and student responsibility

Responsibility stays with the student

From Week 3 onward, coding agents may help produce, explain, test or repair code where the activity permits it. Students remain responsible for experimental design, verification, interpretation, attribution and all claims in their submission.

Earlier labs can permit tutoring or debugging while requiring assessed tasks and written answers to be completed independently. Follow the published lab instructions. Permission in a lab does not imply permission in an exam; consult its examination paper.

Use assistance transparently

  • Explain substantive assistance and distinguish generated or adapted code from your own contribution.
  • Inspect and test the result; do not submit code, calculations or citations you cannot explain or verify.
  • Follow the activity’s restrictions on data uploads, external services, pretrained models and sharing.
  • Do not use a tool to bypass an individual checkpoint or a prohibition on outside assistance.

The general policy covers attribution, honest reporting and reproducibility.

TODO(instructor): Publish any activity-specific permitted-tool lists, disclosure format and assessment evidence, and link the institution’s procedure for misrepresentation. Generated references must be checked; submission details belong to the released assignment.