Google has confirmed that it uses videos on its YouTube platform to train its AI models. The company said it only uses a certain subset for this purpose and respects special agreements with content owners. However, it turns out that most content producers are not aware of this practice.
YouTube videos are used in AI training
Google feeds its AI systems, such as Gemini and its video production tool Veo 3, with YouTube content. A spokesperson for the company said that YouTube content has long been used in product development processes and that this has not changed in the AI era.
Google also claims that various protection mechanisms have been developed to prevent unauthorized use of users’ images and voices. However, experts say there are serious questions about the functioning of the system.
It is not known how many of the approximately 20 billion videos on the platform are used for training purposes. Even if Google uses only 1 percent, this amounts to approximately 2.3 billion minutes of content. This amount is far more than the volume of data that rival companies have access to.
The fact that most of the data was evaluated without user permission raises intellectual property issues for content owners. Some content creators stated that they were not aware that their videos were being used in this way. It was noted that YouTube did not directly inform them that it was training AI with these contents.
In a blog post published in September, Google stated that the content could be used to “enhance the product experience.” However, content creators do not have the option to prevent their videos from being used in AI training.
So what do you think about this? You can share your views with us in the comments section below.
{{user}} {{datetime}}
{{text}}