[key points]

Key points

• •Nvidia is accused of using pirated books to train its AI models, including NeMo Megatron-GPT.
• •The lawsuit, filed by American authors, relies on internal communications suggesting data acquisition via 'shadow' libraries like Anna's Archive and LibGen.
• •Nvidia defends itself by contesting copyright infringement and invoking 'fair use,' arguing that books are transformed into statistical correlations.
• •The court decision, expected after the hearing on April 2, 2026, could have significant repercussions on AI development and language model training practices.