Inteligența artificială și cadrul legislativ actual al dreptului de autor. Studiu de caz CHATGPT
Artificial intelligence and the current copyright legal framework. ChatGPT case study

Summary/Abstract: Lately, several artificial intelligence tools have emerged that have been trained with various texts to learn human language and function to provide answers using human language. One such tool is ChatGPT. The texts used to train these types of tools are not necessarily only texts that are in the public domain, but also texts that are still protected by copyright. This is why, for example, the Authors Guild (a US professional organization for published writers) and The New York Times have sued OpenAI and Microsoft for copyright infringement, with the plaintiffs arguing the tool's training including using their works without obtaining their consent. Regardless of the solution that the court in the United States of America will pronounce, it is interesting to analyse the current legal framework of copyright in Romania and to see if there really is a violation of copyright by these tools, with all the more as we can start from the idea that if, for example, a natural person reads several books, studies, articles, and later offers advice to other people based on the information read, there is no question of copyright infringement. The analysis will consider both the training mode of ChatGPT and the responses provided by ChatGPT, and it should be noted that it is not intended to be an exhaustive analysis, but one that allows certain aspects to be considered to be used in creating a possible future legislative framework.

• Details
• Contents