Previous reports indicate that technology companies including Apple, NVIDIA, and Anthropic have adoptedA dataset created by the nonprofit artificial intelligence research group Eleuther AI，藉此訓練其人工智慧模型，而使用資料涵蓋源自4萬8000個YouTube頻道、超過17萬3000部影片內容的文字內容，而蘋果在稍早向9to5Mac網站Response, referring to itsOpenELM artificial intelligence modelAlthough this dataset is used, it is not used for artificial intelligence application features such as the "Apple Intelligence" service.
In addition, Apple also emphasized that the creation of the OpenELM artificial intelligence model is only for research purposes, and has also opened this artificial intelligence model to the outside world. However, there are currently no plans to create any new version of the OpenELM artificial intelligence model.
In previous explanations, Apple emphasized that its artificial intelligence models are trained based on licensed data, including public data collected through web crawlers and licensed data used for enhanced training of specific features. However, the content trained with the controversial dataset created by Eleuther AI was not used for artificial intelligence application features such as the "Apple Intelligence" service.
Eleuther AI has generated considerable controversy because its dataset, which includes videos shot by famous influencers MrBeast and MKBHD (Marques Brownlee), as well as content from the New York Times, BBC, and ABC News, has been used by technology companies including Apple, NVIDIA, and Anthropic to train their artificial intelligence models.
