Whenever you click on on the Spotify profile of Clever Band Machine you will note a picture of three younger males staring moodily again into the digicam. Their profile confirms that they’re a “British band,” “influenced by the post-punk scene,” and attempting to seize the spirit of bands like The Remedy “whereas carving out their very own distinctive sound.” Whenever you hearken to their music you may be reminded of Pleasure Division’s Ian Curtis.
Should you dig somewhat deeper and examine them on their document label’s web page you will see that that Cameron is the lead singer, and his musical tastes had been formed by the concert events he attended at Nottingham’s Rock Metropolis nightclub. Tyler, the drummer, was certainly impressed by The Remedy, in addition to U2, and The Smiths, whereas guitarist, Antonio, blends his Italian mom’s love of traditional Italian people songs together with his British father’s ardour for The Beatles and The Rolling Stones.
What these profiles don’t say is that Clever Band Machine is just not actual, not less than not within the human sense. And I ought to know, as a result of I created them.
I used a spread of generative synthetic intelligence (GenAI) instruments, in addition to my expertise as an expert songwriter and sound engineer to make their debut album, “Welcome to NTU,” and I launched it on my devoted AI document label, XRMeta Information, in Could 2025.
You may ask why an independently releasing singer-songwriter and music producer like me would create a synthetic band. In addition to being a musician, I’m an tutorial with a background in laptop science, finishing up analysis about how GenAI can be utilized for music.
I had reservations about these instruments and the way they could have an effect on me as a musician. I had heard about varied AI controversies, like “pretend” Drake, and artists like Grimes embracing GenAI in 2023. So, I used to be additionally intrigued by the chances.
Over 100 million folks have tried Suno, an AI music technology platform that may create songs with vocals and instrumentation from easy textual content prompts. Greater than 100 million tracks have been created utilizing the Mubert API, which permits streaming to platforms like YouTube, TikTok, Twitch, and Instagram. And in line with Deezer, 28 % of launched music is absolutely AI-generated.
It was time for me to analyze what these instruments may do. That is the story of how I experimented with GenAI and was reworked from a dance artist to a post-punk delicate rock band.
GenAI Has Modified All the things
In my early days of songwriting, one of many first items of kit I purchased was a Panasonic RQ-2745, a small, slim moveable cassette tape recorder that allowed me to document tough drafts of vocals on an audio cassette tape.
When low cost merchandise just like the Sony cfs-w30 boombox started to include double cassette decks, I may overdub songs and add choruses or devices, like flute or guitar, at residence. If I needed a top quality recording, I needed to e book a recording studio. I grew to become an skilled at splicing tape to take away vocal elements from the tape recording or to repair tape jams.
Reducing and taping grew to become reducing and pasting as I experimented with the very early free digital music sequencers that had been included on a disk I discovered on the duvet of a PC journal. I felt liberated when sequencers like Cubase, Professional Instruments, and Logic allowed high-quality recordings to be produced at residence. This, together with the numerous discount in the price of studio tools, led to the emergence of the bed room producer and the proliferation of the 808 sound. This deep, booming bass line may be heard in hits like “It’s Tough” by RUN DMC, “Emergency Room” by Rihanna, and “Drunk in Love” by Beyoncé.
Digital distribution and social media then paved the way in which for self-releasing impartial artists like me to speak instantly with followers, promote music, and bypass document labels.
But throughout all of those adjustments, musicians nonetheless wanted the abilities and data to create their songs. Like many musicians, I honed my expertise over a number of years, studying to play the guitar, flute, and piano and growing sound-engineering expertise. Even when AI-powered instruments started to be included into digital audio workstations, a musician’s talent and data was nonetheless wanted to make use of these instruments successfully.
Having the ability to create music from textual content prompts modified this.
Not because the introduction of music streaming companies within the late Nineteen Nineties has there been such a dramatic shift in music composition and listening applied sciences. Now non-musicians can create studio-quality music in minutes with out the intensive coaching that I had and with out having to purchase devices or studio tools.
Now anybody can do that. It was time for me to study what these instruments may do.
I usually produce RnB/neo soul, nu-jazz, and dance music, though I can write songs for a number of genres of music. For the experiment, I needed to strive a style that I don’t often produce music for.
I examined about 60 completely different GenAI instruments and platforms. These included standalone instruments that target one process, like MIDI technology (musical knowledge that may be performed again on a keyboard or music sequencer). I additionally tried AI music studios. These platforms have user-friendly interfaces that mix a spread of AI instruments to help lyric, music, picture, and video creation.
Suno and Udio had been two of the very best platforms. They’ll generate songs with advanced vocal melodies and harmonies throughout a spread of genres, with the very best outputs being troublesome to differentiate from what human musicians can create. Each Telisha “Nikki” Jones and music mogul Timbaland are stated to have used Suno to create music for his or her AI-generated artists.
In June 2025, Timbaland introduced the signing of his AI artist TaTa to his devoted AI document label, Stage Zero. In September 2025 Jones was reported to have signed a $3 million cope with Hallwood Media for her AI-generated artist Xania Monet.
On the time of my experiment in March/April 2025, each Suno and Udio had points, comparable to silence gaps, tempo adjustments, inconsistent vocal high quality, and variations in style. Generally the voice may change throughout the track. There was restricted management by way of modifying, and the audio high quality may fluctuate inside a single observe or throughout a sequence of songs.
After attempting a number of GenAI music platforms I made a decision to make use of Udio as a result of high quality of its output and its favorable phrases and situations at the moment. Taking inspiration from pop-rock and post-punk bands like Pleasure Division and The Remedy, I began the journey in the direction of creating a brand new persona.
Utilizing GenAI to supply one or two good songs was fairly easy. Producing an album of 14 songs that sounded as in the event that they had been performed by the identical band was more difficult, notably producing the identical male voice and musical type for every track.
The songs had been both far too related to one another or had different points, such because the voice altering or the devices sounding too completely different. A cautious hearken to the songs in Unfolded by the AI artist Xania Monet will reveal related inconsistencies. For instance, you may hear a distinction within the voice that’s generated for the primary track, “This Aint No Tryout,” in comparison with “Again When Love Was Actual.”
GenAI Can’t Write (Respectable) Lyrics
My first process was to create the lyrics. I generated about 1,000 songs utilizing Udio and located repeated phrases and phrases within the lyrics like “neon,” “whisper,” and “we’re, we’re, we’re,” showing each inside and throughout the 2 person accounts I created. Themes like darkness, shadows, and lightweight had been additionally repeated throughout the lyrics for a major variety of songs.
GenAI simply couldn’t write lyrics with the complexity or playfulness I wanted, so I selected to jot down the lyrics for the album myself and used a semi-autobiographical narrative. This allowed me to take care of a narrative throughout the album; from arriving at Nottingham Trent College and settling into scholar lodging to experiencing college life, graduating and leaving.
I may interweave present affairs just like the closure of Nottingham’s Victoria Centre Market within the track “Goodbye Vicky Market.” I included traces that referenced Nottingham’s historic figures like Alan Sillitoe, who wrote “The Loneliness of the Lengthy Distance Runner,” and the writer D.H. Lawrence within the track, “Books.”
After writing the lyrics I generated the music. There have been points with immediate adherence. I examined prompts of various lengths. In some instances, prompts had been partly or wholly ignored. I’d write a immediate asking for one style and a special style could be produced.
There have been additionally points with the artificial voice saying among the lyrics. For instance, it couldn’t pronounce “NTU” or “Sillitoe,” and I needed to rewrite among the lyrics phonetically or edit the audio to get the right pronunciation for sure phrases.
I relied on my sound engineering expertise, extending the outputs, modifying, mixing, remixing, and manually recording vocals in Cubase to attain a coherent remaining combine. This took a major period of time. Actually, modifying the Udio outputs took a lot time, it will have been simpler to recreate the music myself. I can write a track in 10 minutes, and I generally document myself freestyling lyrics for a whole track instantly in Cubase, so this was irritating.
I encountered related points with immediate adherence when producing photos and video. When utilizing Kling AI to create photos of the band members, I adopted its immediate engineering information. Nevertheless, I needed to generate a whole bunch of photos and edit them with exterior instruments to attain the ultimate band photographs.
Producing video was equally difficult. One technique to create a video is to add a photograph, which turns into the primary body. The remainder of the video is generated primarily based on the immediate. Nevertheless, after I uploaded Cameron’s profile picture to Kling AI, the preliminary frames of the 10-second video resembled him. However by the top of the video, Cameron typically morphed into another person, and this occurred continuously when producing video.
Prompts for digicam directions, comparable to zoom and pan, had been continuously ignored. I additionally needed to edit out scenes with different issues, comparable to the looks of additional fingers or a further leg on the band members.
All this wasn’t low cost both. With 8,000 Kling AI credit at a price of $64.99, I may generate about 40 ten-second movies, however many had been unusable.
Music technology is cheaper. Paying between $24 and $30 for a month-to-month subscription may permit a person to create between 2,000 and three,000 songs, relying on how the credit are used. I used to be very shocked to find how shortly these track credit may be consumed. Each error or track that didn’t swimsuit my style nonetheless value credit.
Finally, after producing hundreds of songs, a whole bunch of photos and movies utilizing instruments like Duck.ai to create the band’s biographies, and spending many hours modifying the outputs; Cameron, Tyler ,and Antonio started to emerge because the band.
AI-generated band Clever Band Machine: Antonio, Cameron, and Tyler. Picture Credit score: Priscilla Angelique-Web page
One thing Sudden Occurred
I’ve all the time been enthusiastic about creating my very own music. As a lot as I like writing songs, the poor royalty payouts I used to be receiving had turn out to be disheartening. A track I recorded in 2001 and launched in 2011 known as Solely Heaven Can Evaluate was streamed about 1 million instances in France throughout 2024 however I solely obtained about £21 in royalties.
Previous to streaming, had my track been downloaded by simply 10,000 folks, I might have been paid about £6,900 (69 pence per obtain). Artists like Kate Nash have raised issues concerning the poor royalty payouts to musicians, citing her £500,000 payout for over 100 million performs of her track “Foundations.”
However as I created the band’s album one thing sudden began to occur. I started to get pleasure from creating music once more. The frustrations with utilizing GenAI was balanced by surprise and curiosity.
At instances Udio was capable of generate vocals that had been so lifelike I may hardly consider they had been created by an AI mannequin. There have been moments after I laughed, after I was actually moved, and even had chills after I heard among the songs.
Lyrics that when lay dormant in a number of lever arch information on my bookshelf started to seek out new life by these generative instruments, permitting me to quickly check them throughout a number of genres.
I made a decision to take this experiment additional.
After fastidiously deciding on a set of songs I had written a few years in the past, I created a brand new persona, Jake Davy Smith. For his 14-track album, known as I’ll Be Proper Right here, which was launched on November 22, 2025, I used Suno’s v5 mannequin to generate studio-quality music that matched my unique imaginative and prescient.
Suno’s intensive modifying instruments permit customers to add vocals, create a canopy track, and edit the music, lyrics, or voice with higher precision than their earlier fashions. This helped me almost recreate my unique songs. The observe “Calling” is an instance of a rock ballad I wrote years in the past, recorded, and didn’t launch.
Conflicting Feelings
Reflecting on this experiment, I discovered myself with conflicting views about utilizing GenAI. These instruments are quick and reasonably priced (in some instances, utterly free). They’ll produce prompt outcomes. I now have instruments that I can use to shortly reimagine my previous songs.
I can use a number of personas to carry my lyrics to life. I’m Priscilla Angelique. I’m Clever Band Machine. I’m Jake Davy Smith. I’m Moombahtman 25, a male African American moombahton artist who combines hip hop with Latin American beats, and I’ve many extra personas.
I’m a “a number of persona musician” or MPM, a time period I’ve created to outline my new musical identification. Musicians having alter egos isn’t new, however GenAI has utterly modified how that is accomplished.
Nevertheless, there’s one other facet to this. Human musicians at the moment are having to compete with algorithms able to producing high-quality music at scale—in addition to with one another.
These instruments are enhancing quickly, and the problems I skilled when utilizing Udio to create the album for Clever Band Machine in March/April 2025 have already been addressed in Suno’s v5 mannequin. It’s now simpler to create a persona with a constant voice. Customers can add their very own songs and in addition create cowl variations of their songs.
Creating the album for Clever Band Machine took about one month and there have been a number of points with attempting to create persistently sounding high-quality, AI-generated songs. I spent hours reviewing hundreds of outputs after which extra time modifying the ultimate set of curated songs in Cubase.
My expertise was very completely different after I created the album for Jake Davy Smith. I used lyrics I had already written, generated between 5 and 20 variations of every track, and spent far much less time modifying them. The method was sooner, nevertheless, there have been nonetheless some points. Modifications in Jake’s voice occurred, although they had been much less frequent and simpler to right. There have been additionally issues with pronunciation, however I may now shortly regenerate the audio. In essence, what had beforehand taken a month now took solely every week.
Moral Points and Knowledge Assortment
But beneath this lies an additional inner battle associated to the info used to coach these AI fashions or, as music journalist Richard Smirke describes it, “the biggest IP theft in human historical past.” It’s this challenge that has made a know-how that should have been celebrated as one of many largest technological achievements in a long time, probably the most contested as a substitute.
Chatbots like ChatGPT, estimated to have almost 1 billion customers worldwide, have been described by the linguist and activist Noam Chomsky as each “marvels of machine studying” and the “banality of evil.” Picture turbines like OpenAI’s DALL-E have additionally come underneath hearth. Critics like Ted Chiang problem whether or not AI could make artwork and different commentators have criticized the dearth of cultural variety in picture technology.
Along with this, in 2024 the UK authorities introduced it was contemplating an exception to copyright regulation that might permit business to make use of copyrighted works for AI coaching with out compensating the creators. This led to protests. Greater than 1,000 musicians launched a silent album known as Is This What We Need in protest in opposition to unauthorized AI coaching. The artists included Kate Bush, Annie Lennox, Damon Albarn, and The Conflict.
Elton John and Paul McCartney additionally voiced their opposition to adjustments in copyright regulation that might profit AI corporations. The thriller about whether or not a band known as The Velvet Sunset was AI-generated added gas to the hearth and sparked additional debate in the course of the summer time of 2025.
But AI corporations have been successful, or not less than partially successful, courtroom instances. In November 2025, Getty Photos “misplaced its declare for secondary infringement of copyright” in opposition to Stability AI. Different AI corporations are making offers, and this consists of Udio and Suno’s latest offers with music corporations. Nevertheless, extra different platforms are rising. Klay.imaginative and prescient is negotiating with the large labels previous to launching, and Soundraw solely makes use of music created in-house for AI coaching.
So GenAI is right here to remain, and musicians might want to adapt. Library music, background music, and music for social media or movie can simply be created with AI. Nevertheless, there are dangers. The danger that related music could also be generated for different customers; the danger that any uploaded songs could also be used for coaching knowledge. Then there’s the danger that these instruments might inadvertently generate one thing that breaches another person’s IP.
A method for musicians to soundly use GenAI is by coaching fashions utilizing their very own knowledge, as YACHT did after they used their again catalog of songs as coaching knowledge for a brand new album. On this approach musicians can have full management over the outputs. That is one thing I will probably be exploring for the following stage of my analysis.
What AI Can’t Do
My transformation has been something however easy. It has been marked by the deep frustration I encountered when initially utilizing these instruments, an ongoing battle about how these instruments are skilled, and moments of real amazement. The albums I created could also be imperfect, however they’re a transparent departure from my common type and present how GenAI can help musical creativity.
Priscilla Angelique-Web page aka Priscilla Angelique, sporting an AI-generated jacket and shirt. Picture Credit score: Priscilla Angelique-Web page
Financially, the albums are unlikely to recoup the price of creating them, as impartial musicians may have a whole bunch of tens of millions of streams to earn a good earnings from music. Even a couple of million streams of the songs will barely cowl the varied charges for music, picture, and video technology of round £140. Merchandise, licensing, sync offers, and different income streams will seemingly stay essential sources of earnings for musicians, whether or not they’re human or AI-generated.
On the authorized facet, one potential approach ahead is for AI corporations to make open-source variations of their fashions freely obtainable for offline use. Some have already got, however for those who haven’t, it appears honest that if they’ve used our knowledge to construct these programs, they need to permit broader entry to the fashions themselves.
New applied sciences may change how music is produced. We’ve gone from clapping to drumming and from utilizing drum machines in recording studios to producing “new” sounds with AI. But now that I’ve accomplished these experiments, I notice that one factor stays the identical.
Whether or not I’m reducing tape utilizing scissors, reducing and pasting in a sequencer like Cubase, or regenerating elements in an AI music studio like Suno, human creativity remains to be a necessary a part of the method. Utilizing GenAI was transformative, but it was my inventive selections that formed the songs, the albums, the avatars for my personas, their biographies, and the general imaginative and prescient. That is one thing that AI can not do – not less than, not for now.
This text is republished from The Dialog underneath a Inventive Commons license. Learn the unique article.