The success of generative artificial intelligence (AI) raises complex legal questions, particularly in relation to copyright. OpenAI and Microsoft, its main investor, are the subject of several lawsuits for violating copyright of works. The two companies are accused of using this content to train OpenAI language models (LLM).
However, as part of a UK House of Lords Digital and Communications Committee investigation into language models, the startup behind ChatGPT admitted that “it would be impossible to train today’s best AI models without using copyrighted materials.” copyright,” Engadget reports.
According to OpenAI, it is currently impossible to escape copyright law, which covers “virtually all forms of human expression, including blog posts, photographs, forum posts, snippets of software code, and government documents.”
OpenAI hides behind “fair use”
Furthermore, “the use of training data from public domain books and drawings created more than a century ago could constitute an interesting experiment, but would not result in AI systems that meet the needs of today’s citizens,” analyzes the startup.
At the same time, OpenAI has attempted to restore its image by insisting on respecting copyright laws despite the accusations leveled against it. “The training of AI models using content publicly available on the Internet is covered by the principle of fair use,” argues the company, referring to this American legal principle according to which certain content protected by the author’s copyright can be used in certain circumstances. .
Furthermore, OpenAI says it is “disappointed” and “surprised” by the accusations of New York Times, who accuses him of using the press articles he published without authorization. “Our conversations with the New York Times It seemed to be progressing constructively until our last communication on December 19. Negotiations focused on a high-value partnership around real-time viewing with attribution in ChatGPT, in which the New York Times you would get a new way to connect with your new and existing readers, and our users would have access to your reports“, laments the AI giant in a blog post in response to the American newspaper.
Source: BFM TV
