DJs from Mars started their career back in the 2000s, and it wasn't so easy in the beginning—the duo had to struggle to make people understand that this was really serious, that DJing was a real job. Thousands of nightclub sets and years of studio work resulted in success. And the word "success" isn't an exaggeration.
A fusion of musical genres in mashups, YouTube videos that amass millions of views, remixes for Sean Paul, Pitbull, Ennio Morricone, tracks and remixes that garner support and inclusion in playlists by renowned DJs like Tiesto, David Guetta, Carl Cox, Chuckie, Oliver Heldens, Lost Frequencies, and others is only a brief (literally!) description of what DJs from Mars have garnered during their career.
This Italian duo is among those DJs who openly incorporate AI in their work. Kill the DJ sat down with the guys to talk about production, the gear they use, and artificial intelligence in DJing.
With AI, the mashup ideas are now unlimited.
Despite the common misconception or merely out of blatant ignorance, the music industry often forgets that AI in music isn't just generative models that steal pop stars' voices and make deepfakes. There are plenty of AI tools that actually help composers, DJs, and sound producers do various tasks—from extracting vocals from a song to mixing and mastering a track.
"The main thing for us (since we produce a lot of mashups and bootlegs) is the acapella extraction. We always struggled to find good-quality acapellas, and now, AI changed the game totally. We can extract good-quality acapella from basically every existing track. The mashup ideas are now unlimited, that helps our studio production, and our set, of course, since we can play any track we want today."
But that's not the only use case for DJs.
We can produce in total freedom, and this is really great.
"Sometimes, we also (let's say) love the bass sound of a new Deadmau5 track, and we want to use that sound in our productions. So we just extract the bass stem, cut a single note, load it into a midi sampler (Simpler in Ableton), and we can immediately play that bass sound with our own bass melody. This is really amazing, so we don’t have to spend hours finding that sound anymore.
"We can use a drum loop from The Beatles in an electronic track. We can have Nile Rodgers from the Chic playing bass in one of our electronic beats. Sampling is the best way to add 'flavours' to your productions. The mix of different worlds creates your own unique style. Now, with AI, it's basically like having the whole music ever created at your disposal 😊
"We remember, before AI, we had an idea, 'let’s make a remix of this track for our sets, it would be great to play it' but then, 'oh no, there’s not a good acapella for thisone, let’s leave this idea behind.' Now, this is not happening anymore, we can produce in total freedom, and this is really great!"
In the 80s, you needed a studio full of expensive hardware. Now, you just need a laptop.
"In the 2030s, you'll probably just need your phone 😊
"We produce 'in the box' with no external hardware. Our favourite VST instrument is Serum, our DAW is Ableton, and for vocal extraction, we use LALAL.AI—it was love at first sight! A suggestion for producers: use an external spectrum analyser—it immediately gives you the idea of what frequencies you’re working on, even when you’re choosing a single snare, clap or kick. That quick visual feedback helps a lot while producing and choosing sounds.
"AI is making things easier, generating a drum loop, sampling a bassline, is now easy and quick. So everyone can do that. But in the end, just like when we entered the computer era, only the best ideas will stand out. So it’s not about the way you do it, it’s about what you do.
"Basically, every DJ is using it for acapella and stems extraction, but things are evolving quickly, so who knows what the next step is gonna be."
If your songs are bad, you'll only be noticed as a bad songwriter who used AI.
Dozens of artists and content creators now make songs with pop stars' vocals, put their voices in their tracks, and garner thousands of listens and views. One of the cases is the notorious "fake Drake"—an anonymous TikTok creator who made a viral AI song with Drake's and The Weeknd's voices. The song got millions of streams on Spotify and views on YouTube and TikTok before being taken down by all platforms. Guys don't seem to be impressed that much and think that if you're bad at it, AI isn't going to make you better.
We don’t think you can make a good song just asking AI 'make a good song.'
"AI is just a tool that needs to be used by a human. It simply allows you to have technical possibilities that you would never had before, but in the end, you are leading the way it’s been used. You can use Drake’s voice to create a brand-new song, okay. But if you can write a good song, that’s going to be noticed and liked. If your songs are bad, you'll only be noticed as a bad songwriter who used AI 😊 So the evolution will be driven by natural selection, as it always is, when a new technology comes out."
Mashup producers should be able to sample, and royalties should be paid to the original master owners. It should be the same for AI.
Copyright concerns are on the rise as AI tech becomes more accessible to the public. When everyone can create a song with anyone's voice or backtrack with no legal consequences whatsoever, the rights of an original owner are so easily violated.
For now, all we know is that, AI-generated works that don't have "human authorship" are not copyright-protected, as per the US Copyright Office. The reasoning is that "copyright should only apply to original works reflecting human creativity."
And this neverending copyright debate isn't exclusive to AI-generated or AI-assisted music. When it comes to sampling and making mashups, similar questions arise.
"For us, as mashup producers, copyrights should be almost free. Or at least you should be able to sample, and royalties should be paid to the original master owners. That’s what major labels do lately with mashups on YouTube—they allow the mashup, and they earn from it. So everyone’s happy with this. It’s a win-win. It should be the same for AI."
If someone prompted an AI tool to create a remix in our style, we would be honored.
We asked DJs from Mars if they were worried about AI as a tool that could harm their own reputation. Say, if someone creates a mashup that sounds like their own one and mimics their style. Guys aren't concerned.
"If someone prompted an AI tool to create a remix in our style and uploaded it everywhere, we would be honoured. And we’re sure it will never sound like an actual remix produced by us. It would be slightly different, and maybe it could suggest a different direction to us.
"You can’t stop ideas, this never happened, and it never will. When sampling was going out of control in the 90s (remember the first albums of De La Soul, or Beastie Boys), they changed the copyright rules. They weren't able to stop sampling; they just made some rules to make sure the original master’s owner was getting paid. That’s the way it should be with AI. AI and samplers have a really similar concept, and, as you can see, sampling wasn’t stopped.
"But where’s the line between 'same' and 'similar' and 'inspired by'? Macklemore’s voice sounds a bit like Eminem's, right? Should he be stopped from singing? This debate would take forever, and we’ll never find a solution 😊 but generally, for us, art should be 100% free. Of course, we see something risky in deepfakes—you can upload a video where Joe Biden declares war on Italy! But we think humans will always recognise when technology is bad. Or at least, they should, but you can’t stop it.
"We don’t think you can make a good song just asking AI 'make a good song.' But you can definitely ask to create a melody, and then use your own taste to make it great. So use it in a wise way, let your mind and taste choose the final direction."
