Grimes fans can now create the Indie pop musician’s voice with her new artificial intelligence vocal software Elf.Tech. The singer announced the program’s launch via Twitter on April 30.

The singer tweeted a thread outlining how the software works and encouraged musicians to use her vocals in their work. The website allows users to record or upload audio and uses an AI model to make the vocals sound like Grimes. The program was created with the AI engine Triniti.

Grimes is open to musicians using AI model’s vocals in commercial works and is offering to split 50% of the royalties. She claims there are “no legal bindings” because she isn’t signed to a label.

However, Grimes explained that she would take down any works that promote hate speech or extreme violence.

“[I] don’t wanna be responsible for a Nazi anthem unless it’s somehow in jest a la producers I guess,” she tweeted.

Fans are using the software to generate everything from a cover of A-ha’s “Take on Me” to original compositions inspired by the musician’s early discography.

Elf.Tech isn’t the only AI project Grimes has launched. In 2020, she created a series of AI–generated lullabies with the algorithm-based music app Endel and sang about AI’s potential on her 2018 single “We Appreciate Power.”

On April 28, she shared her thoughts on AI and copyright in the music industry during a keynote speech at the International Music Summit in Ibiza.

“Art is a conversation that’s come before us,” Grimes said. “Intertwining it with the ego is a modern concept. The music industry has been defined by lawyers, and that strangles creativity.”

Grimes’ manager, Daouda Leonard, told Rolling Stone’s “Music Now” podcast that the singer launched the software because she “sees music as a gift to be shared, not hoarded.”

Both Leonard and Grimes compared the AI voice clones to fanfiction.

“That’s a beautiful relationship between the consumer and the fan and the creators of that IP,” Leonard said. “And so that’s what we’re jumping on. That’s what we’re excited about.”

Grimes isn’t the only musician experimenting with AI voice models. Singer Holly Herndon’s website Holly+ provides a similar service where uploaded audio converts the vocals to sound like Herndon.

“Vocal deepfakes are here to stay,” Herndon said in a statement to Pitchfork. “A balance needs to be found between protecting artists and encouraging people to experiment with a new and exciting technology. That is why we are running this experiment in communal voice ownership.”

Although Herndon and Grimes have been receptive to the technology, other musicians such as Drake have pushed back against AI-generated music.

Universal Music Group, Drake’s record label, took down the viral song “Heart on My Sleeve” from streaming platforms and cited copyright violation because it used AI models to emulate Drake and the Weeknd’s voices.

Although the song isn’t available to stream, numerous uploads survive on YouTube. A channel called Snapshot Edits uploaded an album consisting of AI generated Drake songs with features from AI Kanye West and AI Tyler, the Creator. As of May 8, the video has 325,000 views and 15,000 likes.

Drake disapproved of an AI generated cover of Bronx rapper Ice Spice’s breakout hit “Munch” and said it was “the final straw” on his Instagram story.

Singer-songwriter Nick Cave also criticized fans for using the AI text generator ChatGPT to write lyrics in his style.

“The apocalypse is well on its way,” he said. “This song sucks.”

Grimes also announced two new songs, “Music for Machines” and “I Wanna Be Software.” She didn’t specify when the songs will be released.