Distinguishing between fraud and plagiarism in the context of generative artificial intelligence use

Cases involving fraud, plagiarism, and/or non-compliance with recommendations concerning the use of generative artificial intelligence (GAI) are subject to Article 23 of the Faculty's General Study Regulations.

The general definition of plagiarism at the University of Geneva is available at this link. Situations of plagiarism involving students are specified in Directive Ref. 0008 of October 25, 2011.

What is considered plagiarism?

Plagiarism consists of presenting the work or ideas of others as one's own, without mentioning the source.

Examples:

• Copying a text without citing the source

• Using a text generated by GAI without indicating this

• Rephrasing ideas from an external source without attribution

• Translating or rewriting a passage using GAI (e.g., DeepL) without indicating it.

The use of a GAI tool to generate content, without any proper declaration of said use, is considered plagiarism.

What is considered fraud?

Fraud is intentional deception aimed at obtaining an unfair advantage (credit, grade, funding, etc.).

Examples:

• Submitting work generated by GAI and presenting it as entirely personal, despite this being prohibited

• Using GAI to invent scientific data

• Signing a declaration of originality when the content was generated by GAI.

Fraud involves a deliberate intent to deceive and constitutes a violation of academic regulations and is contrary to the Code of Ethics and Professional Conduct.

How to distinguish plagiarism from fraud in the context of GAI?