AI-generated music has become common enough in electronic dance music circles that some producers have started acting as unofficial fact-checkers, calling out tracks they suspect were built by algorithms rather than artists. The trend has grown alongside more capable audio generation platforms, and it has split the EDM scene between listeners who cannot tell the difference and producers who insist they can hear it instantly.
Among the most vocal is Max Harris, a 26-year-old producer who performs as H4RRIS. Harris has built a following by posting videos that dissect suspected AI tracks, breaking down the audio artefacts he believes give them away. He argues that generative tools produce what he calls “a kind of decoy art form”, one that mimics the surface of human creativity without any of the intent behind it.
Why the callouts started
Harris says his objection isn’t really about the technology itself, but about what he sees as a shortcut around the creative process. His own workflow, built around Ableton Live, a Novation Launchkey 49 controller, Ableton Push 3, Launchpad X, a pair of analogue synthesizers and software like Serum, Diva and Kontakt, involves hundreds of small decisions per track. Each tweak to a melody, a mix or a transition is, in his words, a deliberate choice aimed at “evoking a specific kind of idea or emotion”.
That contrast is central to his argument. He believes AI-generated music strips out exactly the decision-making that defines a producer’s style, replacing it with statistical guesswork drawn from existing songs. For Harris and producers like him, calling out suspected fakes isn’t just gatekeeping. It’s an attempt to protect the credibility of a genre where audiences increasingly can’t tell what they’re hearing was made by a person.
The tells producers listen for
According to Harris, AI-generated tracks tend to share a handful of audio fingerprints once a listener knows what to listen for. He points to a faint, persistent hiss running underneath many generated songs, which he attributes to how these models often start from a block of white noise and then reconstruct waveforms by referencing patterns pulled from training data.
Another giveaway, he says, is vocals and melodic elements stuttering in sync with each other. Generative models still struggle to fully separate the different layers of a song, sometimes treating vocals, drums and synths as a single blended instrument rather than distinct parts. To an experienced producer, that kind of glitch reads as something no human would deliberately choose to leave in a finished mix.
Visual cues sometimes reinforce the suspicion too. Harris has flagged accompanying videos where character animations glitch or fingers appear to vanish mid-frame, along with an overly polished, uncanny visual gloss that often accompanies AI-assisted content across platforms.
Suno and the copyright question
Much of the frustration in the scene has centred on Suno, the AI music generation platform that producers like Nihil Young, a 39-year-old Italian turntablist-turned-producer, blame for a recent spike in suspect uploads. Young, whose posts about the issue partly inspired Harris to start his own callout videos, argues that some users are feeding copyrighted tracks by established artists into the platform and asking it to generate “remixed” output.
He points to a cover of Madonna’s “Like a Prayer” that was later updated with AI credits following backlash, as evidence of how easily existing songs can be repurposed without the original artist’s consent. Young’s concern isn’t limited to famous names. He argues the same process could just as easily be applied to any independent producer’s back catalogue, feeding years of original work into a generator to produce new tracks with none of the original creative input attached.
Neither Suno nor the artists named in these disputes, including MANSA and the duo behind “Take Me (To The Moon)”, have confirmed publicly whether AI tools were used in the disputed tracks. Harris has been upfront that some of his suspicions rest on pattern recognition rather than proof, which is itself part of the wider problem: without transparency from artists or platforms, EDM fans are left relying on producers’ ears rather than confirmed facts.
What it means for listeners and the industry
The rise of this informal detective work says as much about the platforms enabling AI-generated music as it does about the producers pushing back against it. As generative audio tools keep improving, the artefacts that give away a synthetic track, like the stuttering or background hiss Harris describes, are likely to become harder to detect, not easier.
That puts pressure on streaming services and social platforms to build clearer disclosure standards, rather than leaving verification to individual musicians combing through waveforms. It also mirrors a broader pattern playing out well beyond music, as generative systems and local AI training tools becoming more accessible make it easier for anyone to produce convincing content with minimal expertise.
For now, the EDM scene’s response has been grassroots rather than institutional: producers publicly dissecting tracks, fans debating the evidence, and platforms occasionally updating credits after backlash rather than proactively. Whether that model scales as AI-generated music becomes harder to distinguish from the real thing remains an open question, one that echoes similar debates about AI reshaping how people interact with digital platforms more broadly. What’s clear is that trust in the scene now depends partly on producers willing to do the listening work that platforms haven’t yet automated a fix for.
