YouTube is giving select users a chance to test new AI capabilities for music creation, generating enthusiasm and questions around the emerging tech.
YouTube is inviting creators to experiment with AI tools that can conjure full songs from text prompts or remix tracks in new styles, powered by datasets including seven major artists. But how will these technologies impact music and where do artists’ rights fit in?
What you need to know:
- Dream Track allows generating 30-second snippets using the voices of stars like Demi Lovato and Charli XCX
- A second tool called Music AI can automatically convert genres or generate instrumentals from vocal clips
- Concerns remain around training AI without permission and how artists will profit as the space evolves
YouTube’s new Dream Track experiment debuted this week. The AI music feature uses the voice datasets of nine pop and R&B stars to synthesize audio based on text descriptions like “upbeat” or “melancholy.” One example generated a “T-Pain style” track from basic input.
Artists lent vocals to “explore opportunities AI could offer,” according to Charli XCX, while Demi Lovato views it as “pushing boundaries.” However, Dream Track currently only produces short snippets and other details like license terms are unclear.
A second new tool called Music AI can perform broader manipulation like changing a song’s genre. Google DeepMind CEO Demis Hassabis says converting a clip to country based on its whistle or generating a full instrumental from gibberish shows AI’s “multimodal” capabilities.
Benefits and challenges of AI music tools
Proponents argue these AI music technologies could streamline creation and spark new forms of expression. YouTuber Lyor Cohen believes they’ll let users “seamlessly turn thoughts into tracks.” However, training datasets raise issues around permission and compensation. Some artists now seek to prohibit their work from fueling commercial AI.
As with other generative media, assurances are needed that new music tools respect copyrights while connecting fans to artists in novel ways. YouTube pledges transparency through audio watermarking, but oversight will grow more complex as model abilities increase. Early discussion focuses on direct revenue sharing, though wider economic impacts are still emerging.
Youtube wants to compete with TikTok
YouTube’s new experiments also target the short-form video giant TikTok, where music both enables viral sharing and forms a major revenue stream. The company says AI effects and remixing functions could enhance clips in similar viral, sampling styles.
This competitive angle matters since TikTok largely pioneered integration of AI into the creative process at scale. But as more programs emulate its formula, guarding artistic interests will require nuanced policy. Overall, YouTube’s debut tools represent an optimistic industry test case for navigating AI’s role in reshaping music.
In sum, AI music creation sparks debate but also opportunities. As both technologies and partnerships continue evolving rapidly, open dialogue between creators and platforms seems key to widespread benefits. Only by addressing complex rights issues proactively can the human–AI relationship in music fulfill its collaborative promise.
