Recent developments in two prominent AI-related legal disputes have brought renewed attention to copyright issues, particularly concerning Meta and OpenAI. In a case against Meta, newly disclosed court documents have ignited discussions about the company's alleged use of pirated e-books from Library Genesis (LibGen) for training its Llama AI models. There are allegations that Meta employees attempted to erase copyright identifiers from various works, with one internal document proposing the removal of specific terms associated with copyrights. Communications among staff members also indicated a competitive eagerness to surpass rivals like OpenAI's GPT-4 and a dismissal of French competitor Mistral as insignificant. Mark Zuckerberg's deposition further emphasized the complexity surrounding the use of pirated content. He acknowledged that while such practices might initially appear dubious, Meta's teams take various nuances into account. Despite claiming ignorance of LibGen, he expressed caution about collaborating with entities that overtly engage in illegal activities. Additionally, internal documents suggest that Meta executives realized the implications of incorporating LibGen materials, warning of potential repercussions under the EU AI Act. In parallel, OpenAI faces allegations from The New York Times, which asserts that the tech firm improperly used its content to train AI models without consent. The argument put forth highlights a greater concern about copyright infringement occurring without adequate tracing mechanisms in place. Amidst these legal challenges, partnerships between tech firms and publishers continue to evolve, as evidenced by Axios collaborating with OpenAI to fund local newsrooms and the Associated Press teaming up with Google. Conversely, Apple has paused its use of AI news alerts due to inaccuracies, while regulatory bodies like the FTC and U.K.'s Competition and Markets Authority are increasing scrutiny on AI practices within the industry.
This news is summarized and processed by the IP Topics artificial intelligence algorithm.
Read the full article on the original webpage: https://digiday.com/media/ai-briefing-copyright-battles-bring-meta-and-openai-datasets-under-the-microscope/
The cover image belongs to the source website and is used as an integral part of the summary of the reference article.