Apple refutes claims that it trained Apple Intelligence with stolen data - here's what we know

Apple refutes claims that it trained Apple Intelligence with stolen data - here's what we know

In response to reports that it used stolen YouTube video data for Apple Intelligence training, Apple clarified its position. The company stated that while it has used the data in the past, it has never used it to train Apple Intelligence.

It was recently revealed that an AI lab called EleutherAI had harvested subtitles from YouTube videos without the creators' consent. This data was then combined with data from Wikipedia, the British Parliament, and Enron staff emails and added to a data set called “the Pile.”

Apple was accused of using the data from the Pile to train Apple Intelligence along with companies like Nvidia and Salesforce. Recently, Apple stated that it had used this data set in the past, but only to train the OpenELM model it released in April. Apple later confirmed to Apple Insider that the OpenELM model does not have AI or machine learning capabilities and was created to contribute to the research community.

Apple stated that the OpenELM model was not intended to be used for Apple Intelligence and that it has no plans to build a new version of the OpenELM model. Apple has repeatedly claimed that it uses only ethical sources for its artificial intelligence projects, including paying millions of dollars to publishers and licensing images from photo library companies.

Apple Intelligence aims to revolutionize Apple products and will make significant changes to Siri, including the ability to maintain context in conversations and make them feel more natural. Added to this is the news that Siri will become smart enough to understand and take action on apps that are on the screen at any given time, including several new in-app actions. However, not all of these features will be released at once, and some will not be seen until at least 2025.

AI production has skyrocketed recently, raising concerns about how data is collected. The news that many companies have used Pile is concerning, but it is good to see that Apple is mostly focused on dealing with ethically sourced data.

For a complete breakdown of everything coming in the next generation of Apple software, see Apple's full WWDC presentation.

Categories