Hazen E. Griffin reviews an AI-generated album.
“AI-generated” has become something of a mark of Cain in the artistic world; a moniker somewhere in the ballpark of “spiritually Israeli” and “chopped” on the persona non gratae scale. So it goes. The very concept of AI music intrigues me, because I seriously struggle to imagine a use case for it. If you’re reading this, I doubt you knowingly listen to AI-generated music, if at all, and I don’t expect to be finding many people who unironically do. I’ve yet to hear an AI-generated song that I could feasibly hear being played in most any public venue; at the club, in a bar or pub, being blasted out of someone’s phone on public transport... I can feasibly imagine the kind of person who regularly views AI-generated images or art, or somebody who watches AI-generated YouTube Shorts or wears a t-shirt with an AI-generated graphic. All of these people exist, and you’ve probably met them or are related to them familially. But who are the people regularly listening to machine music? Do they live among us, flying just under the radar and living double lives that inevitably end every day with a voice-controlled candle-lit evening spent reading AI-generated fanfiction and listening to the latest computerized lo-fi hip-hop beat? It baffles the mind.
The Guardian reported in September of last year that Spotify removed over 75 million AI-generated tracks from their platform in 2025 alone, some of which had hundreds of thousands of streams. I figured the best thing to do would go straight to the source on this stuff -- who would know better about where to find AI-generated music than AI itself?
I asked Elon Musk’s Grok, which will generate just about anything, and DeepSeek, the open-source Chinese alternative to ChatGPT. Both directed me to Reddit and Discord, which seemed like places where fans of AI-generated music would congregate. On Reddit I found the greatest minds of the AI music industry congregating and sharing their ideas for the future of the genre, like Panos_pro0’s suggestion of “[E]urovision plus AI sings [sic] and characters”, or LaBaricala’s attempt to revive the career of 1950s country artist Patsy Cline using the magic of computerized generation.
An issue I quickly ran into is that none of the people prompting the generation of AI music seemed interested in the album as a medium. The closest you could get was playlists of aesthetically similar songs, usually in textural-based styles like post-rock or trance, with the occasional text-to-speech poem set to a sultry guitar lick. Writing one’s own lyrics was touted as a point of pride. The emphasis seemed to be on the self-generation of music tailored exactly to your interests as opposed to any grand attempt to make the next Dark Side of the Moon-style concept album using the magic of ChatGPT.
But who are the people regularly listening to machine music? Do they live among us, flying just under the radar and living double lives that inevitably end every day with a voice-controlled candle-lit evening spent reading AI-generated fanfiction and listening to the latest computerized lo-fi hip-hop beat?
Finding a genuine album to review was a serious challenge. Eventually I settled on the 2025 album “Dog Dog Dog” by AI-generated “band” DogLand, which bills itself as a post-punk outfit of three dogs probably from France. The group boasts a discography of songs mostly centering on the experience of being a dog, with singles such as “I have hair, everywhere” and “My Leash”. Unlike most of the songs I heard browsing through the AI Music subreddit, DogLand’s music sounded relatively polished, and didn’t put forth a glaring admission of being AI-generated in its sound; other tracks I sampled sounded like the audio equivalent of an AI-generated image of a person with six fingers.
The title track, “Dog Dog Dog”, is an electronic-indie rock synthesis heavy on the 80s-style drum machine and dream-pop guitar. Not much in the way of lyrics (as is expected of AI), there nevertheless remains this tinge of auto-tuning over the soft indie-rocker vocals that really seals the 2011 electronica vibe... my immediate thought was that it sounded like something that would be used in an Adult Swim commercial. There’s clear inspiration here from Empire of the Sun, one of those bands you’d hear on the radio in the early 2010s. It’s hardly worth mentioning the next track, which re-uses the instrumentals from “Dog Dog Dog” to playtest the idea of using real lyrics on the album.
“Echoes of Barking” is DogLand’s staggering seven-minute foray into 90s grunge, complete with female vocals somewhere between Portishead and Sonic Youth’s Kim Gordon (I wonder which of the three dogs does the singing… or do they rotate?). There’s serious artifacts of “Black Hole Sun” and other such grunge anthems peppered throughout the track itself, but the instrumental quality has seriously fallen off from the prior two tracks -- this one sounds like it was made in GarageBand, or roughly sampled from ripped YouTube videos. The last three or so minutes of the song just repeat, so this is definitely one to cut off early.
In the same vein, “I Always Run” is an emo post-punk shoegaze blend that draws heavily from Have a Nice Life’s “Trespassers W”. The general schema of the song is in line with the greats of the genre; Joy Division, Slowdive, My Bloody Valentine, et al. … the drum machine is back in full force, and whichever one of the three dogs that sings like that asshole from Empire of the Sun returns singing generally canine lyrics like “I run, I run, I run, I run… I run always!”
“I Dream” hits the niche of the recent UK post-punk revival, and appears to be based entirely off the works of the band Shame. There’s whiffs of IDLES that shine through at times, and even a bizarre triangle section towards the end of the track, which seems to have actual mastering done on it. This is one of the only songs on the album that actually focuses on the vocals, which appear to take the forefront against a complex but forgettable instrumental track. This one was actually listenable, which is unfortunate because it is the shortest song on the album.
“Kiss Me Like a Dog” is a swan song for the 60s psychedelic rock scene, and blends Beatles-esque vocals with instrumentals that clearly draw on a number of sources and nails the magic of none of them. This entire album is like cheap wine, and trying to find every band or song that’s had their identifiable elements fed into the digital woodchipper and regurgitated as a very well-stitched Frankenstein is like trying to find notes of chocolate and hickory in something that was clearly made to be consumed en masse by teenagers trying to get drunk. The song sounds like the memory of a mediocre Beatles cover band.
By this point, I was losing patience with the album. The lo-fi commercial filler tracks are nearly five minutes long and the dog-themed lyrics about running and playing fetch have become less cute and more eye-rolling. Lyrics like “Where is my ball? Who am I without my ball? Have you seen it?” really hammer home the sinking feeling that the fun is over -- was there a contractual obligation to get to 50 minutes? Did the guy generating the album run out of genres that he liked?
The last two tracks, “Free as a Dog” and “We are Dogs Twins”, were previously released by DogLand as singles in the months leading up the album. Both are largely forgettable, and I didn’t really feel the need to finish either of them.
The album as a whole strikes me as an audio version of leaked journal entries. These are clearly songs that were tailor-made to fit one person’s musical taste – the linked Spotify profile of the person behind DogLand confirms that their music taste includes nearly all of the bands I’ve mentioned in this review – but nevertheless released to the public in the form of an album strung together only by the motif of dog-themed lyrics. We laugh now, but this is the future. Your Daily Mix, generated For You by the algorithm, is the first step in the direction of songs generated For You, and all of these things come bearing the banner of convenience above all at the price of a subscription fee. Make no mistake about it -- this is economic music.