Meta’s AI development push has led to significant controversy as the company faces multiple copyright challenges over its Llama model, including a recent case in France where French publishers have accused Meta of using their books without authorization.
The issues stem from the same drive to build a rival AI model to OpenAI’s ChatGPT. According to reports, Meta CEO Mark Zuckerberg approved the use of copyrighted material to build out its language model, despite concerns over “ethical” concerns around taking intellectual property from artists.
In France, publishers and authors have launched a lawsuit against Meta, accusing the tech giant of using their books to train its generative artificial intelligence model without permission. Similarly, in the US, a significant legal challenge has been filed against Meta over unauthorized use of copyright-protected material to train its Llama model.
The problem with Meta’s approach is that most content from users’ apps isn’t helpful for building its AI model due to factors like deletion of older posts and inconsistent writing styles. As such, the company relied on pirated books as a source of new data.
This has raised concerns about potential lawsuits around the world, especially if initial cases lead to compensation deals for impacted authors. If legal precedent can be established, it’s likely that every publishing house will seek traces of their own works, leading to major penalties for Meta moving forward.
OpenAI also faces similar challenges and investigations over unauthorized use of copyrighted material. The company’s success relies on having the best data sources, which is why large language models will eventually win out based on their reference set.
Source: https://www.socialmediatoday.com/news/meta-faces-lawsuit-france-copyright-protected-works-ai-llm/742369