“Restore the truth, put an end to the false texts.
This is the slogan chosen for his computer program by Jérémy Goffin, an 18-year-old student following a first year in cybersecurity bachelor's degree at ESAIP, the engineering school of Saint-Barthélemy-d'Anjou (Maine-et- Loire).
Targeted ?
ChatGPT, the conversational robot developed by the Californian start-up OpenAI.
The student has developed StendhalGTP, an application to identify whether a text was written by artificial intelligence (AI) or not, reports
Ouest-France
.
The student designed a website stendhalgpt.fr, on which he explains his approach.
"StendhalGPT is an application dedicated to recognizing texts generated or reformulated by artificial intelligence, using the diversity of the lexical field used", can we read.
The operation of the program is simple, just copy and paste the text into an appropriate field and press "verify".
Only in English
The tool, which is still in development, only works with English text.
But the student will soon work out a version to check the French texts.
It is by studying the structure of the text that StendhalGPT can issue probabilities that the text has been generated or reformulated by an artificial intelligence.
Company
Lyon: Half of the students of a master's degree caught cheating thanks to artificial intelligence
By the Web
ChatGPT: The artificial intelligence tool fascinates as much as it worries
Angers
Pays de la Loire
Artificial Intelligence (AI)