Earlier this week, a survey revealed in detail how Apple and other tech giants use YouTube subtitles to train their AI models, including over 170,000 video segments from channels such as MKBHD and Mr. Beast. Subsequently, Apple used this dataset to train its open-source OpenELM model, which was released in April this year. However, Apple has now confirmed that OpenELM does not support any of its AI or machine learning functionalities, including Apple Intelligence. Apple stated that the creation of the OpenELM model was intended to contribute to the research community and promote the development of large-scale open-source language models. In the past, Apple's researchers have described OpenELM as the "state-of-the-art open-source language model." According to Apple, OpenELM is only used for research purposes and is not used to support its Apple Intelligence functionalities. The model has been released as an open-source project and is widely available, including on Apple's machine learning research website. Since OpenELM is not used as part of Apple Intelligence, this means that the "YouTube subtitles" dataset is also not used to support Apple Intelligence. Apple has previously stated that the Apple Intelligence model is trained on "authorized data," which includes selected data to enhance specific functionalities and publicly available data collected by the company's web crawlers. Finally, Apple also stated that the company has no plans to develop a new version of the OpenELM model. As reported earlier this week by Wired magazine, several companies, including Apple, Anthropic, and NVIDIA, have used this "YouTube subtitles" dataset to train their AI models. The dataset is part of a larger-scale dataset called "The Pile," released by the non-profit organization EleutherAI.