Antiplagiarism publishes a new algorithm to identify and tag content generated by ChatGPT

CTOnews.com, May 16, Antiplagiarism recently released a new algorithm for identifying and tagging fragments of papers generated by the AI system, based on years of research in the field of natural language processing (NLP).

At present, the new algorithm is open to Antiplagiarism paying users and can identify the content of papers created by GPT-2, GPT-3 and ChatGPT language models.

After Antiplagiarism detects traces of AI generation, the system marks the relevant material, and the status of the document is changed to "suspicious".

According to Yuri Chekhovich, executive director of Antiplagiarism, neural networks "always leave a trail". In addition, in his view, there are differences between all generated text and manual writing, allowing you to identify documents created by artificial intelligence.

CTOnews.com Note: the Antiplagiarism website uses the world's largest open database to detect the existence of digital plagiarism detection programs.