Top News

Google launched the 'Lyria' model within Gemini in February of this year. It is capable of generating songs based on prompts. Shortly after its release, a lawsuit was filed against Google, alleging that the company used songs uploaded to YouTube to train this model without providing any compensation. Google has now surprised everyone by filing its response in court—a move seen as a major blow to content creators. Let’s look at what Google has said regarding this matter.

Google's Response Comes as a Surprise

The company informed the court that whenever an artist uploads a song to YouTube, they grant Google a license to use it for AI training purposes. Google stated that YouTube's terms of service authorize the use of music uploaded to the platform for training AI models. This implies that the moment a song is uploaded to YouTube, Google obtains permission to use it for its own operations.

Did Google Train AI Models Using These Songs?

It is important to note that Google did not explicitly state that it trains its AI models using songs uploaded to YouTube. Google's counsel argues that the allegations made in the lawsuit are incorrect. Furthermore, even if the allegations were assumed to be true, there would be no grounds for a lawsuit because the uploader has already granted a license for the use of the uploaded content. If this argument is accepted, Google would gain the full right to train its AI systems using songs uploaded to YouTube.

Google's Approach Differs from Other Companies

Google's approach in this matter differs from that of other companies facing allegations of copyright infringement. In the cases involving Suno, Udio, and Anthropic, the companies argued that training AI on copyrighted material is not wrong, as it leads to the creation of systems capable of driving large-scale transformation. They claim that their use of copyrighted content constitutes "fair use" and, therefore, does not amount to infringement.