In the mid-2010s, when I was in high school, the genre known as lo-fi beats was like a brain cleanse for all my time spent mainlining Minecraft and YouTube. The music was tender, with drums sizzling like kitchen pans and trumpet that zigzagged over the soundscape’s open pasture. Perfectly timed samples leapt out like flickers of a memory. Only a few channels soared back then; the mixes all had whimsical titles and background visuals that simulated the feeling of sitting indoors, cocooned in a duvet, while rain lightly falls outside.
Fast-forward to now, and the scene has putrefied into a wasteland of the percussive undead. The YouTube search results have warped into an apparent AI breeding ground, crammed with hourlong mixes full of soporific dreck. The channels have similar names and cartoonish Kawaii imagery. Even the comments of the videos, which have millions of plays, brim with what look like fake conversations—pseudoymous accounts prattling on about how the music helped them unlock their true potential in life. Multiple channels repeat the same sentence structure like, “I don’t want much! I just want the person reading this to be healthy, happy, and loved!,” suggesting they’re AI-generated. “To whoever reads this comment: you are a wonderful person and I love you,” writes one channel called Coffee Time, which specializes in videos with titles that read more like Airbnb listings than music: “Happy Morning Jazz | Elegant Bossa Nova In Luxury Cliffside Cabin Overlooking the Sea for Relaxation.” It’s the theme song for the dead internet theory, a rave-yard of zombified AI agents chilling out to death.
Longtime heads who fell in love with lo-fi beats and delicately craft it themselves are starting to panic. Obsessives on the Reddit forum for lo-fi beats, which banned AI submissions late last year, recently despaired about how the scene has been “overtaken” and “lost its soul.” Some artists report losing significant opportunities and having to switch careers because of the genre’s downturn; others are paranoid, unable to discern the real from the hoax.
Mia Eden, a 23-year-old from Manchester, England, is one of a slew of lo-fi beatmakers who explicitly state that their videos are “made by real people” and free of AI. She runs a channel called Lofi Louis, inspired by the name of her friend’s pet rabbit, and records under the alias Rosia! She started making the music in 2021 after stumbling on channels like the infamous, perpetually studying Lofi Girl. Eden dug deeper and found the underground lo-fi scene, a community of artists who happily shared advice, made lo-fi beat-themed podcasts, and collaborated on compilations. Years ago, Eden earned anywhere from $500 to $1,500 a month from her songs getting slots on DSP editorial playlists; now, after the AI boom, that’s mostly dried up—Spotify’s lo-fi playlists spill with viral songs from suspicious profiles with no descriptions. Eden first noticed a dumpster’s worth of AI cover art, and then people full-sending the grift with AI music. She says she’s been tricked by some of her favorite channels starting to sneakily integrate AI.
“Previously, you could stream a track on Spotify or Apple and almost be certain you followed them on Instagram or spoke to them on Discord because the community was so tight-knit,” she says. “Now, it feels so nameless—where this could be an artist that maybe doesn’t like to show face, or it’s a computer. You can’t always distinguish now, and I’d say it’s over half [AI].”
Alex Reade, a 32-year-old from the United Kingdom, thinks the genre is “the most uninviting it’s ever been,” choked in a swamp of shit: listless, derivative beats and pure AI slop. He came to the genre around the turn of the decade, and was intrigued by the idea of trying to infuse things like post-rock into lo-fi. While working overnight shifts at a tech company’s store in an emptied-out mall, he’d balance out the spectral creepiness of the space with chill lo-fi beats. Like Eden, he’s seen a sharp dropoff in streams under his alias Project AER; a skyrocketing count of two million listeners per month a few years ago has sunk to 420,000 today. Spotify’s fickle playlist placements and lo-fi’s enshittification has freaked him out. “There’s so much anxiety around what to do as artists,” he told me. “I’m trying to find any other means so I can take that reliance that I have on lo-fi out of my life because it causes me a lot of stress.”
Many people credit lo-fi beats to the mothbitten jazz-hop of J Dilla and Nujabes. It’s a slightly misleading term, since the sound isn’t always low-fidelity, and the phrase “lo-fi” already has many other meanings in music. It goes back to indie rock in the ’90s with one-man noise-makers and fuzzy bands like Guided by Voices, who made a deliberately imperfect, clangorous sound—partly as a reaction to the pristine quality of the CD era. “Lo-fi beats” coasts on the vaguest of associations, and there was never an ideological motive behind it. But its success online might partly come from it being the antithesis of so much ultraglossy, hi-fi music in the 2010s, from hyperpop to mainstream trap and pop. Lo-fi beats are frayed and slipshod, with a dustiness that carries the hallucination of lowkey intimacy, like you’re overhearing someone practicing their drumming.
The scene proper didn’t really crystallize until the mid-2010s, when Lofi Girl and other online hotspots cropped up. The music was unofficially christened as “lo-fi beats to study to,” taken from the clickbait title of so many YouTube mixes that curated soothing yet stimulating instrumental music. It became a macro-genre catchall term for anything vaguely chill, jazzy, wistful, playlistable: capacious enough to cover both J Dilla and cosmic IDM. The scene had a slew of mini-stars—producers like eevee and potsu, who’s maybe most known for making the petal-soft beat that XXXTentacion later hijacked for his massively popular “Jocelyn Flores.”
The dreamiest beats offered a kind of divine comfort, from sublime bossa nova soundtracks for Pokémon games that don’t exist to the bittersweet beats sampling Shiloh Dynasty, the enigmatic vocalist who is basically the scene’s Imogen Heap. Listening to the Flume-ified tripstream of eery’s “hardly” instantly takes me back to the days when I first heard it as a teen—it inspired me to crack FL Studio and try to create my own instrumental mischief. (Thankfully, I pivoted to writing.)
AI has annexed the lo-fi scene for a hodgepodge of reasons: The lack of vocals, which typically gives away robo fraudulence, make it easier to infiltrate; the music’s association with aimless, unfocused listening—vibe music before vibe became a buzzword—means people aren’t paying as much attention to what’s real and what’s not; the fixation on fantastical, Studio Ghibli–core visuals, which image generators can vomit up with ease. Take Mewy Cat Lofi’s “Relaxing Lofi for Study Time,” which lures passersby with adorable animations of Pusheen before, five hours into the 12-hour mix, they realize the same chord progression has seemingly repeated 100 times. YouTube offers creators the option to disclose if their videos use AI, but they only require disclosure for some things, such as “altering a famous car chase scene to include a celebrity who wasn’t in the original movie.” A few viral mixes have “Altered or synthetic content” warnings in the description, but many don’t.
“It’s a travesty, truly,” groans Dreamwave, a 26-year-old from Washington whose channel was one of the earliest and biggest lo-fi archives, and who has never used AI for his carefully curated videos. He describes the arc of his channel as a slow, soul-crushing descent, from a thriving community of lo-fi lovers to minimal views on his uploads. He says it’d take him a month to sculpt a three-hour mix of lo-fi music—discovering and blending the sweetest tunes, asking for permission—while autobot channels can do it with a few clicks. “With AI, you can just come up with some ridiculous chord progression and then turn it into an ambient track where there’s just not a lot going on,” he explains. “You can almost loop it over and over and then upload that. You can upload three hour compilations every day.” He believes he’s lost millions of views because his sparse uploads have been deprioritized in the search results. “It really pisses me off to see anything AI-generated getting so many views. It enrages me.”
“The oversaturation caused by AI-generated music is very real,” adds Berkkan B., the manager of Lofi Records, the label spearheaded by Lofi Girl. “It’s flooding the platforms, and unless streaming services implement some kind of regulation, which we hope they will, this will inevitably dilute the presence and visibility of real artists.” While Berkkan believes AI “can be a powerful tool” to do things like “enhance workflows” and “refine ideas,” they say that everything on the Lofi Girl channel comes from human composers and designers.
I spoke with a pseudonymous creator who uses AI to power four separate channels: jazz music, meditation, rain sounds, and Lofi Tone Art, the latter of which has amassed over 10 million cumulative views. The clips often show looping gifs of murky cityspaces and rain-soaked cabins that offer a “quiet sanctuary in the storm.” They told me they use AI to create everything: ChatGPT for descriptions; other unspecified software (likely Udio or Suno, the most popular programs that generate music from text prompts) for the audio. “To be honest, I think at this stage, AI still struggles to outperform real music due to its high error rate,” they told me in an email. “However, when used as a supportive tool, it can be incredibly helpful. When selecting songs, I usually listen to them repeatedly to ensure they sound smooth and don’t overpower the mood.”
Detractors might argue that lo-fi beats was always mercenary music engineered and optimized to hook sad bois with no taste, and good riddance. But to the aficionados, it’s sacred music—the musical madeleine for a generation that listened to it through the highs and lows of their adolescence. Dreamwave says every upload to his channel represents a memory in his life. Reade sounds jubilant as he expounds upon his love for the somnambulant churn of Philanthrope and Sleepy Fish’s “Space Cadet,” a track he claims he “could literally put on repeat and just listen to forever.”
For these lo-fi fanatics, the music came with the bonus of a lovely community. Eden described the scene as something like a virtual neighborhood, with specific comps supporting women beatmakers and online friends she’s since met in real-life, like the founders of the Portuguese label Salad Day Records. Reade recalls being in a cluster of Twitter chats, along with a Discord channel called “Lofi Backstage” that’s a who’s-who of “peak era” lo-fi GOATs. Dreamwave used to play Rocket League with other YouTuber-archivists, and relished the raw thoughts people left on his videos. “A lot of the comments on my channel are usually people going like, ‘Hey, I remember when I was listening to this song with my ex-girlfriend seven years ago, and I just wanted to comment and say that we broke up.’ You can’t find that in an AI-generated video.”
Many of these artists are fighting back, writing screeds against AI and commissioning art from illustrators whose livelihoods are being threatened. There’s a hint of futility in their voices as they wonder about what lo-fi will look like in a few years, but also a plucky determination, like they’re bedroom-producer Davids battling big bad Gol-AI-ith. It’s likely that as AI advances and makes reasonable facsimiles of even more genres, there’ll be a reverse push for realness—akin to listeners obsessing over vinyl or obscure formats, the human touch could become a boutique feature, like raw milk at the farmer’s market. Lo-AI hasn’t won the battle yet; it means too much to these people. “When things get so bleak with lo-fi or just the world, right, it’s very easy to enter a state of nihilism. Like, why should I do any of it, when it’s all so fruitless?” Reade said. “You do it for yourself. That’s the core thing with art and music for me.”