Katarzyna Petru
ElevenLabs – a company that has risen to fame for generating voice using AI – has just announced the debut of a new service Eleven Music. This tool enables users to generate their own music based on a simple description in natural language. There is no need to be knowledgeable about composition – simply type something like: “create smooth jazz with a 60s vibe, strong lyrics, and an atmosphere perfect for a Friday afternoon” – and the artificial intelligence will create a track with vocals and instrumentation in a matter of minutes. For content creators, filmmakers, app owners, or small businesses, this could be a monumental change – less cost, fewer formalities, more creative control.
Playing Open Cards with Record Labels
To avoid falling into the same problems as its competitors, ElevenLabs chose the path of legalisation and collaboration. The company has already signed agreements with Merlin Network, which represents independent labels, and Kobalt Music Group, a well-known player in the world of copyright management and music publishing. For now, they are determining which catalogs will be used to train the model, but the mere fact that such discussions are taking place is a signal: ElevenLabs wants to operate sensibly. Importantly – the company emphasises that it does not use music from the major labels (UMG, Sony, Warner), but hopes to soon negotiate terms of cooperation with them as well. This is crucial because without permission to use the data, training AI on someone else's recordings is a legal roulette.
Security First
Of course, this does not mean that ElevenLabs is unaware of the threats. The company immediately implemented safeguards to prevent the generation of songs featuring specific artist names, quotes from albums, or lyrics that may be illegal, offensive, or inciting violence. The aim is to avoid instances of "deepfake hits" featuring Drake’s voice or unauthorized songs masquerading as existing performers. ElevenLabs wants to create a tool that offers users freedom but does not cross ethical and legal boundaries.
Who is all this for?
The new service aims at creators who care about music for advertisements, applications, films, or games – that is, where you usually have to pay for a licence or hire a composer. According to ElevenLabs, already 20 companies and creators are testing the system. It is not known exactly who, but examples of applications include television productions, games, meditation applications, or even fitness ones. Industries such as automotive, telecommunications, or creative agencies could also quickly jump on this bandwagon – especially since AI can replace expensive stock music for literally "a fraction of the price."
Creators versus Technology
However, not everything looks like paradise. Organisations protecting the rights of creators – such as ASCAP – warn that AI can be a threat if it does not operate fairly. Creating models based on other people's works without consent poses a risk not only of lawsuits but also of undermining the incomes of entire professional groups. As the head of ASCAP, Elizabeth Matthews, said, AI can be an innovation, but only if it respects the rights of the people who have created music for years. And it is hard to disagree with this – because while technology races ahead, creativity should still be grounded in respect for human labour.
Customer Delight or Receiver Anger?
There is also the issue of image. Even if the music generated by AI is legal and works flawlessly, it may still be a problem for many receivers and customers. Analysts such as Mike Proulx from Forrester warn of potential backlashes – social support for “real artists” and concern for human jobs are very strong emotions today. Companies that start to massively use AI instead of employing people may face a backlash – in terms of image, emotionally, and in some cases, financially.
