Policies & Guidelines

NTU Guidelines on Academic Integrity

CodeCraft: Precision in Snippet Generation and Syntax Validation

NTU Staff Only

Use of AI Detector Tools

Research shows that the use of AI detector tools should be used with caution due to the following reasons. 

Unreliable Detection:

  • Frequent false positives (human text flagged as AI) and false negatives (AI text missed).
  • Detectors rely on patterns that are not unique to AI-generated text.
  • Easy for students to bypass detectors using minor text modifications or paraphrasing tools.
  • The mixing of human and AI-generated training data further complicates accurate detection.

Discrimination:

  • Bias against non-native writing patterns (e.g. ESL/EFL students).

Undermining Educational Goals:

  • Detectors lead to false accusations and hinder teaching of future-relevant skills.

Bauschard argues that educators should shift focus from unreliable AI writing detectors to training students and teachers on responsible AI tool usage, preparing students for real-world applications.


NTU Examples

CodeCraft: Precision in Snippet Generation and Syntax Validation

NTU Staff Only

Other Curated Examples

A recent review (Moorhouse et al, 2023) examined publicly available guidelines from 23 of the top 50 universities worldwide ( links to guidelines) on the use of GenAI tools in student assessments. 

Curated examples:

If you (or your student) are submitting papers to academic journals for publication, do check their AI use policy first. A comprehensive list of such policies can be found here.