Like it or loathe it, A.I. is here to stay. Chandler Shortlidge looks at how this is going to play out for DJs.
If 2018 encapsulated any feeling, it was probably bleak. The news cycle seemed like one never-ending reason to fear the future. And stories about the impending rise of Artificial Intelligence probably didn’t do much to dissuade those feelings.
If you believe the headlines, and I do, AI and automation are poised to disrupt massive sectors of the economy—from trucking to retail to healthcare. These industries employ many millions of people, and the possibility of a robot takeover has pushed some politicians and thinkers to discuss distributing a Universal Basic Income as a way to offset the eventual loss of wages facing these workers. Though to really get a sense of where AI is probably headed, and where it already is, we need to look at how its affecting fields traditionally thought of as off limits to automation. Like Instagram influencers.
Yes, even Instagram models risk losing out to lines of code. As the New York Times pointed out in an article on fembots, one such influencer already exists. Lil Miquela is a bundle of CGI contradictions. She’s “woke” in the sense that she supports Black Lives Matter and the Innocence Project, while simultaneously not having any conscious or consciousness at all. She’s completely made up, a digital rendering of today’s hottest beauty standards, posing in real clothes and backgrounds. None of that matters to her 1.5 million followers, who comment by the hundreds that they “love her” as she gripes about how much she hates being “back at work.” “Huge meeting coming up has me feeling like I’m just playing at being An Adult™️ because it’ll never really happen,” she says. “Send good vibes? Please?” So #relatable.
While I won’t shed too many tears over the collapse of the influencer market—okay, no tears will be shed—Lil Miquela’s existence comes at a time when automation encroaches on a more near and dear topic: art. In painting, one portrait created by AI sold for $432,000. Many soon asked, is it art? I think the more important question is, does it matter? If it sells, and people like it, the answer is probably no. But by and large, paintings like this still rely on human input. “Humans are deeply involved with every aspect of the creation and training of today’s AI technologies, and this will continue to be true tomorrow and for the foreseeable future,” Jessica Fjeld, assistant director of the Cyberlaw Clinic at Harvard Law School, told Artsy.
Though how much input is needed varies, and likely won’t remain constant. Take AICAN. Dubbed “The biggest artistic achievement of the year” by Artsy. AICAN was developed by Ahmed Elgammal, the director of the Art and AI Laboratory at Rutgers University, who wrote an algorithm that essentially synthesized 80,0000 images from Western art. Once the algorithm was off and running, it independently produced images novel enough to pass a special Turing test—the computer generated images were rated more likely to be made by a human than human-made art in a double blind study. Or as Artsy put it: “Watson had learned how to paint like Picasso.”
A.I...will never get drunk, doesn’t need to eat or take a break, won’t make mistakes, won’t complain, and won’t ever get caught huffing powders from under the booth..
So what could this mean for music? We got our first glimpse back in 2016 with the pop song “Daddy’s Car.” For that experiment, researchers from Sony used AI to analyse the melodies and harmonies of 13,000 songs, then told the program to write something that sounded similar to the Beatles—though any style could be input. The software wrote the melodies, and a human composed the song and wrote the lyrics. So while “Daddy’s Car” still relied heavily on human input, there’s no reason to think that with an algorithm like the one used for AICAN, that would need to stay true forever. One could easily imagine a program that would analyse the entire history of techno or house to create something new and innovative; something that when listened to by humans, sounds “more human” than human creations.
Similarly, it isn’t hard to imagine software that could analyze every available Jeff Mills set in order to faithfully reproduce a Jeff Mills performance. Hologram post-mortem performances aside, this type of technology would likely have the greatest impact on smaller performers. Simply feed in the top-rated 10,000 deep house mixes from SoundCloud, et voilà, the perfect beach bar DJ is born. It will never get drunk, doesn’t need to eat or take a break, won’t make mistakes, won’t complain, and won’t ever get caught huffing powders from under the booth. You don’t even need to pay AI. Suffice it to say, pub DJs might be in real trouble if AI DJ technology ever becomes a cheaper alternative. But even club DJs might be out of a job. After all, how much do most partygoers really care who their DJ is on a night out? I’m not talking about real fans. Think office workers having a big one, or students hitting the town. If AI is available and cheaper, it will get the gig. Hell, some clubs might even tout AI as the new “futuristic” reason to visit.
What’s harder to imagine is AI DJ software successfully reading and reacting to the crowd. Though it’s not impossible. Take Luka, a company specialises in building AI-powered software like chatbots.
The company’s latest invention is Replika, which learns to become more like you the longer you speak to it, meaning one day your family could have a “replica” of you to talk to long after you’re dead (like Black Mirror). Existential quandaries aside, that type of learning software could surely be programmed into AI DJ algorithms. Sensors placed around the venue could read and analyze inputs like temperature, humidity, noise and movement to learn the optimal settings for any song or style. It wouldn’t be perfect at first, but who is? Given time, it’s hard to believe the machine wouldn’t begin to “read” the crowd with greater and greater accuracy.
So yes, AI is probably coming to steal your DJ gigs. But that’s no reason to completely fear the AI takeover. While there will be a massive shift in how things are done, new opportunities will also likely be created. Human DJs will have to up their game to stand out against the robots, just like they did when digital replaced vinyl. Tech savvy DJs can even work with AI, using its code to better understand the audience’s wants and needs. This applies to producers too, who can use AI to build better and more complex songs with less effort on the part of the human. Like it or not, this is the future we’re building. Human and machine working together to make roads safer, retail more efficient, and music and art more beautiful. What a time to be alive.
Chandler Shortlidge is a dance music journalist based in Berlin. Follow him on Twitter.