The Compilatio AI detector reveals whether a text has been produced by a generative Artificial Intelligence or by a human.
Our detector is itself an AI trained to spot the predictability and markers specific to content generated by AIs such as OpenAI's ChatGPT or Google's Bard.
So, if a student uses a text-generating AI to correct or reformulate his content, will he be detected by the Compilatio AI detector?
- Case 1: student uses ChatGPT or equivalent for proofreading
In this case, the student asks the AI to reread his or her writing in order to correct mistakes.
The basic structure, the words and their arrangement, are therefore initially those of the student himself. They therefore bear no markers specific to a text written by AI.
➡️ Our detection tool therefore objectively considers the text as if written by a human. In this case, ChatGPT can be seen as a writing tool.
- Case 2: the student uses ChatGPT or equivalent for reformulation
In this case, the student asks the AI to rework his initial text. Depending on the level of reformulation required and applied, the student's text can be significantly transformed.
➡️ The reformulations can then become the words of the AI and will be more potentially detected as AI by the Compilatio detector.
This article has been translated automatically. If you have any suggestions for improvement, contact us.