Voice Models and Consent: Grimes and the Open AI Voice Experiment
In April 2023 the Canadian musician Grimes publicly invited fans to create songs using an AI model of her voice and offered to split earnings from successful tracks 50/50, later backing an online tool that let creators transform their own vocals into her sound. At a time when unauthorised AI vocal imitations of famous singers were spreading online and being removed, this was an unusual move: an artist choosing to open her voice under clear conditions instead of fighting every copy. The experiment became a reference point in debates about consent, credit and income in AI music.
Here is what happened, why it mattered and what musicians, marketers and students can learn from it.
The background: imitation without permission
Earlier in 2023, songs that imitated well-known singers with AI voice models went viral. Many listeners were impressed, yet the artists had not agreed, and platforms removed several tracks after complaints from rights holders. The technology had moved faster than the rules, and a common question appeared: should an artist's voice be treated as something that belongs to them, even when no recording is copied directly?
The experiment: an open door with house rules
Grimes took the opposite approach to blocking. The idea had a few key parts.
Permission in advance
Creators did not need to ask individually. The invitation itself served as consent, within published guidelines.
Shared earnings
If a track using the voice earned money, the artist expected a share, presented as an even split.
Credit and transparency
Songs were meant to be identified as using the AI voice rather than pretending to be official releases.
Limits
The artist reserved the right to reject or take down content that conflicted with her values, for example highly offensive lyrics.
Why the model attracted attention
The approach turned a threat into a collaboration and a marketing story. Fans became co-creators, the artist's name spread through hundreds of new songs, and the debate itself generated press coverage. It also tested whether a voice could function like a licensed instrument, available to anyone who follows the terms.
The open questions
The experiment did not settle every issue, and many artists chose not to follow it.
- Quality control: when anyone can release songs in your voice, some will be poor or off-brand.
- Identity: fans may struggle to tell official work from fan-made tracks.
- Fairness: not every musician has the audience to benefit from shared earnings.
- Law: rules on voice rights differ between countries and continue to develop.
Lessons for music marketing
The case shows how consent can be built into a product strategy instead of added after a problem.
- State clearly what is allowed and what is not.
- Make credit easy, with a standard label or tag.
- Offer a fair, simple reward to collaborators.
- Keep the right to remove harmful content.
- Treat the community as partners whose creativity adds value to the brand.
These principles apply far beyond AI. A teacher sharing practice tracks, a band releasing stems for remix contests or a composer licensing a theme for school concerts can use the same clarity.
Lessons for learners
For students experimenting with AI vocals or instrument models, the takeaway is respect. Use tools whose terms you understand, avoid imitating real people who have not agreed, and label your work honestly. Most importantly, keep building your own skills. An AI voice can sing a melody, but someone still has to write it, choose its harmony and decide how it should feel. Pianists who can harmonise a tune, accompany a singer and shape a phrase will always have something to offer any collaboration, human or digital.
A wider industry shift
Since the experiment, labels, publishers and technology companies have explored licensed voice models, opt-in programmes and detection tools. Whatever form future rules take, the core idea that artists should have a say in how their identity is used has gained wide support.
FAQ
Q: What did Grimes offer fans in 2023?
A: Permission to use an AI model of her voice in new songs, with earnings from successful tracks shared evenly.
Q: Was this the same as copying her recordings?
A: No. Fans created new songs; the AI transformed or generated vocals in her style rather than reusing existing records.
Q: Can I use any famous singer's voice in AI songs?
A: Not without permission. Unauthorised imitations are often removed and may raise legal issues.
Q: What can a small musician learn from this?
A: Clear rules, fair credit and a welcoming community can turn fan creativity into positive attention.
Develop your own sound with TIGA ALPHA
TIGA ALPHA is an AI piano teaching app built by a Thai team, with lessons in Thai, English and Mandarin. Its CEO, Teacher Aum, has 20 years of experience in piano teaching, and that music teaching background shapes every lesson, while piano teaching with AI gives you feedback each time you play piano. Use it as an online piano lessons app for short daily sessions, follow it as a complete AI online piano course, or combine it with a piano course or piano learning course led by a teacher. Whether you want to learn piano from zero, learn piano online after a long break, learn music alongside your children or simply play music for fun, this online piano teaching application keeps your next step clear. New users get a 7-day free Premium trial, then 799 THB/month. Open the piano teaching app below and pick a first lesson that fits your week.
- Download and start: TIGA ALPHA landing page
- In-person lessons in Thailand: TIGA Studio
- About the company: tigaalpha.com
Technology can copy a timbre, but phrasing, timing and intention still come from the player.