A shocking new report released recently spoke about big tech companies using YouTube’s data for the sake of training their AI models, without consent.
Amongst those in the list was leading iPhone maker Apple who was shunned by others for stealing data without consent and compensation. A whopping 170k videos from the app were said to be used including those belonging to big creators like MrBeast.
It was also highlighted how the Cupertino firm then made use of the dataset for training open-source ELM models that were rolled out in April of this year.
But Apple has opted to fight back against those claims, with a new statement shared with 9to5Mac on how its OpenELM does not power any AI-related features such as its much talked about Apple Intelligence.
The tech giant added how the new OpenELM model is a great way of helping the whole research ecosystem and how it would take openELM model development to the next level by outsourcing the tech.
We’ve seen in the past how researchers from Apple went on to describe how these models were state-of-the-art tools.
As per Apple, the purpose of creating the models was solely to facilitate researchers and never to power any of its product lineup including Apple Intelligence. Its publishing was done through open source and today, it’s up for grabs for the masses and can be found on the company’s own machine learning website page online.
Therefore, since the OpenELM is not incorporated into Apple Intelligence, it makes sense as to how the app’s subtitles are not utilized to power Apple Intelligence. We’ve seen how Apple has previously spoken about in detail how it chooses to be transparent about its AI model training process.
It’s done through licensed data only where a lot is used solely to better certain features while some pieces of data were taken from the web that were available for public use including those it’s web-crawler could scrap.
What about the future of Apple’s OpenELM model? The company says it has no future plans so far to build bigger and better versions.
Another report by Wired highlighted how top tech giants such as Apple, Anthropic, and NVIDIA were also a part of the tech giant gang who scrapped subtitles from video streaming giant YouTube without consent for AI training purposes.
Image: DIW-Aigen
Read next:
Amongst those in the list was leading iPhone maker Apple who was shunned by others for stealing data without consent and compensation. A whopping 170k videos from the app were said to be used including those belonging to big creators like MrBeast.
It was also highlighted how the Cupertino firm then made use of the dataset for training open-source ELM models that were rolled out in April of this year.
But Apple has opted to fight back against those claims, with a new statement shared with 9to5Mac on how its OpenELM does not power any AI-related features such as its much talked about Apple Intelligence.
The tech giant added how the new OpenELM model is a great way of helping the whole research ecosystem and how it would take openELM model development to the next level by outsourcing the tech.
We’ve seen in the past how researchers from Apple went on to describe how these models were state-of-the-art tools.
As per Apple, the purpose of creating the models was solely to facilitate researchers and never to power any of its product lineup including Apple Intelligence. Its publishing was done through open source and today, it’s up for grabs for the masses and can be found on the company’s own machine learning website page online.
Therefore, since the OpenELM is not incorporated into Apple Intelligence, it makes sense as to how the app’s subtitles are not utilized to power Apple Intelligence. We’ve seen how Apple has previously spoken about in detail how it chooses to be transparent about its AI model training process.
It’s done through licensed data only where a lot is used solely to better certain features while some pieces of data were taken from the web that were available for public use including those it’s web-crawler could scrap.
What about the future of Apple’s OpenELM model? The company says it has no future plans so far to build bigger and better versions.
Another report by Wired highlighted how top tech giants such as Apple, Anthropic, and NVIDIA were also a part of the tech giant gang who scrapped subtitles from video streaming giant YouTube without consent for AI training purposes.
Image: DIW-Aigen
Read next:
