Apple and Salesforce respond to YouTube video complaints

Recently, allegations have emerged that tech giants such as Apple and Salesforce are using content from YouTube videos to train AI models without consent, sparking a major debate about the ethics and legality of using data in AI training.

Claims

The report found that companies including Apple and Salesforce were using millions of YouTube video captions to power their AI capabilities. These captions were part of a larger dataset known as a “pile,” which also included a variety of data sources, from European Parliament minutes to a collection of Enron emails.

Apple's response

In response to these allegations, Apple clarified that the specific AI model mentioned, OpenELM, is purely for research purposes and has not been integrated into any commercial products. Apple emphasized that its AI models, including those for upcoming Apple Intelligence features, rely on ethically sourced and licensed data and information collected by its proprietary AppleBot crawler.

Salesforce's Position

Salesforce, which is also involved through its subsidiary Slack, noted that its data usage policy is entirely internal and aimed at powering AI capabilities without violating privacy norms. The company remains committed to ethical data use, with an emphasis on transparency and user consent.

Ethical considerations

The use of public data for AI training is a grey area that continues to challenge the boundaries of privacy and copyright law. Companies have asserted the legality of using such data for non-commercial research purposes, but the impact on content creators remains a matter of debate.

As AI technology advances, the need for clearer regulation and ethical guidelines is becoming increasingly apparent. Both Apple and Salesforce have taken steps to address these concerns, committing to greater transparency and adherence to ethical standards in their future AI developments.