Искусственный интеллект на музыкальном фронте: как технологии стирают грань между творчеством и алгоритмами Headline: Artificial Intelligence on the Music Front: How Technology Blurs the Line Between Creativity and Algorithms

A study by Deezer–Ipsos reveals that 97% of listeners are unable to distinguish between songs created by artificial intelligence and those written by humans, according to a report by Reuters.

The survey, which included 9,000 respondents from eight countries such as the United States, the United Kingdom, and France, found that 73% of participants would like to see transparency regarding the use of AI in music creation. Additionally, 71% expressed surprise at their inability to tell apart human-made tracks from synthetic ones.

Deezer, a music streaming service with 9.7 million paid subscribers, reports that it sees more than 50,000 AI-generated tracks uploaded daily, accounting for about one-third of all new content, a rise from 18% in April.

The platform has introduced labeling for AI tracks and has removed them from editorial playlists and algorithmic recommendations.

«We firmly believe that creativity is a human endeavor that deserves protection,» stated Deezer’s CEO Alexis Lanturnier. He noted the challenges of implementing a differentiated payment structure for AI-generated music, while also addressing the elimination of fraudulent streaming from royalty calculations.

In early 2025, controversy erupted over the band The Velvet Sundown, which had attracted millions of listeners on Spotify, revealing that their music was synthesized. Another concern is the training of neural networks on songs without obtaining the necessary licenses.

A Munich district court ruled that the startup OpenAI violated German copyright laws by training its models on the lyrics of Herbert Grönemeyer and other artists, including the hits «Maenner» and «Bochum.»

The case was brought forth by the German copyright society GEMA, composed of composers, poets, and publishers. Presiding Judge Elke Schwager ordered OpenAI to pay compensation for using copyrighted materials, although the specific amount was not disclosed.

OpenAI claims that its language models do not store or replicate specific training data, reflecting instead the knowledge acquired from the aggregate information. Since the output is generated after inputting a prompt, users are deemed responsible for it, according to OpenAI.

However, the judge opined that the memorization in LLMs and subsequent reproduction of song lyrics constitutes copyright infringement.

The court’s ruling could set a precedent in Europe regarding the use of copyrighted materials for training AI models. «The internet is not a self-service store; it consists of human creative achievements, not free templates. Today, we have established a precedent that protects and clarifies authors’ rights: even operators of AI tools like ChatGPT must comply with copyright laws,» stated GEMA’s CEO Tobias Holzmüller.

An OpenAI representative noted that the company disagrees with the ruling and will explore options for next steps. «The ruling pertains to a limited set of song lyrics and does not affect millions of people, companies, and developers in Germany who use our technology daily,» he said.

In September, Suno unveiled the «world’s first» generative DAW called Suno Studio, which «radically rethinks the music creation process.»