A new lawsuit has been filed in the United States by authors, alleging that Apple used books written and authored by them for the purpose of training artificial intelligence models. This was done without permission or licensing, thus making the training of the models illegal by Apple.
Apple is said to have used a dataset called Books3 which includes a number of book data illegally, without the permission of the authors or publishers. Apple was found to have used the dataset in training its OpenELM model.
Apple had previously published the OpenELM model on Hugging Face openly, allowing the public to review the source code, in addition to showing the dataset it used.
For now, it is still in the early stages of filing the lawsuit. It is not yet known when and what the final trial process will be. Recently, Anthropic settled the lawsuit by agreeing to pay $1.5 billion to the authors, showing a compensation value of around $3000 for 500,000 authors that were pirated for the purpose of training AI models.
