🤖📚 Authors Challenge Claims on Anthropic’s $1.5B Settlement
Anthropic’s $1.5 billion copyright settlement is facing new problems after some authors said publishers and literary agents were making unexpected claims on their payments.
The settlement covers nearly 500,000 books that were used to train AI models. Authors may receive around $3,000 for each work, but the payment depends on who owns the publishing rights.
Some writers say publishers are asking for money from books whose rights had already reverted to the authors. Others say certain publishers are requesting 100% of payments when they may only be entitled to half.
Experts say the problem may be caused by poor record-keeping rather than intentional wrongdoing, but the number of similar complaints suggests the issue could be more widespread.
Meta has introduced a new pricing option for its AI model, Muse Spark, which is designed for coding and AI agents.
The company is offering a strong incentive: users can get a major discount if they allow Meta to retain their prompts and AI outputs. Meta can then use this information to improve future AI models.
This means Meta could leverage real user interactions to make its AI systems better. In return, developers can access the model at a much lower price.
However, this deal comes with a privacy question. Developers and companies often work with sensitive information, including private code, customer data, and internal business details. For some users, protecting this information may be more important than getting a cheaper AI service.
Would you let an AI company use your prompts in exchange for a huge discount? 🤔