The late music industry lecturer Ed Montano wrote in 2010 that “the DJ will always have a role in club culture”. More than a decade later, technological advances have fully digitised the DJ’s tools, prompting the idea that the DJ themselves could soon be digitised. An automated, artificially intelligent DJ could be an easily achievable cost reduction for venues— so why hasn’t the DJ been automated?
The late music industry lecturer Ed Montano wrote in 2010 that “the DJ will always have a role in club culture”. More than a decade later, technological advances have fully digitised the DJ’s tools, prompting the idea that the DJ themselves could soon be digitised. Currently, dissemination of club culture remains dependent on the hospitality industry (which is rife with profit and labour optimisation) to hold events. An automated, artificially intelligent DJ could be an easily achievable cost reduction for venues— so why hasn’t the DJ been automated?
At its core, the act of DJing breaks down into two components: selection and beatmatching. Track selections are a key way of conveying capital through music, both by the physical labour of searching for the perfect song and the economic capital required to purchase it. From this search, DJs reconstitute the musical zeitgeist in order to entertain an audience. This audience is especially enthralled by a seamless stream of music, wherein breaks between songs are unnoticeable. This involves a technique known as beatmatching, which ensures the music being played is matched in tempo and phase. Through a combination of quantitative beatmatching and qualitative song selection, the DJ hopes to form cohesion between the two through their own personal style.
From a technical standpoint, AI is already well on the way to replicating this process of track collation and mixing. DJ programs can determine timbre and tempo for gigabytes worth of songs in the same time it takes a human to do just a handful. Automatic mixing and transitions are already implemented on certain electronic music playlists on Spotify too. However, AI will always struggle to utilise one of humanity’s most powerful inventions: irony. DJs can find success not just through genuine unironic cohesion, but also through the subversion of the current musical zeitgeist with the occasional trance classic or Vengaboys remix. For an AI to adequately read a crowd correctly, extensive monitoring via facial recognition and other surveillance techniques would likely be required. Clearly, these methods of quantifying people may obstruct the club experience as we currently know it. Club goers may not feel comfortable being surveilled so intensely for an experience that could easily be replicated by genuine human perception instead.
Another aspect keeping the humans in business is the mystery of sound causality in DJ performance. DJs exist in a subsection of musical performance where the sounds perceived by an audience may not match with what they can actually see. Compare this with a guitarist, whose actions and gestures give clear clues as to what sound they will produce. From vision alone we can assume:
- Timbre (electric/acoustic, speed of strumming)
- Sequencing (frequency of hand position changing)
- Pitch (position of hand on the neck)
- Texture (hand shape dictating polyphony or monophony)
Even with turntable DJs, who require a physical mastery due to the tangibility of the vinyl they use, sound causality is still elusive to the audience. The music, while physically there in front of the DJ, is still unreadable to the audience. With contemporary DJing, the physical actions are much more minute and thus more difficult to notice. Without specific knowledge of the equipment used and its functionality, DJ performances can seem like sheer magic. The CDJ player (which is similar to a record player but with USBs instead of records) has largely replaced the mechanical turntable, and thus removed any physical hints the audience could deduce. Furthermore, DJ staging often places the CDJs outside of the audience’s view, preventing them from viewing the DJs’ skills. Instead, the DJ stands tall over the audience as they work. It can be argued that the mystery of causality would also be present in an AI DJ. The workings of machines also require proficiency in a code language in order to be understood, similar to the proficiency in CDJs needed to make sense of a DJ’s physical movements. However, there would be absolutely no way for an audience to see any trace of a digitised DJ, bringing the mystery of causality to a new peak.
DJs exist as mediators of pleasure, it is their job to perform the ritual and our job to acknowledge them. The audience regards the working DJ as a visible symbol of their own labour and staging often intentionally places DJs’ torsos and heads in full view. Subsequently, we as an audience face them despite not being made privy to their workings. Over time, sound systems have also naturally shifted to accommodate crowds wanting to see the DJ (or at least their top half) perform. People are seemingly drawn to their presence, as well as the collective attention and work they provide throughout the night. It’s fair to say this emotional connection would be absent in an AI DJ.
Most examples the masses would conjure of DJs are actually fully live performances. Daft Punk’s Alive 2007 Tour, The Chemical Brothers, Deadmau5— these performances utilise extensive visual automation not to correct but to add on to the preexisting oracular spectacle. The withheld physical presence of the DJ, the source of cohesion, is tastefully obscured by extravaganza. Mystique still persists on the headliner stage, as the visual has become a crucial element of DJ performance especially at festivals and large scale events. Apart from adding an ‘Instagrammable’ element to a performance (a topic worthy of an article in itself), human DJs working in conjunction with technology amplify mystique seamlessly. An automated DJ visualised by an automated visual would be no better than watching a video.
Like many artists in our contemporary society however, DJs have their work constantly devalued in the name of ‘exposure’. In a scene inundated with fresh faces all the time, tech investors simply do not see a need to automate a group of people already working for a pittance. Often they have to give away their skills for free to music blogs and radio shows in the hope of garnering more work. An AI would undermine the human collective struggle heard through the music, as well as the community formed as a result. Innovation and invention of new techniques would come to a standstill, as an AI would simply regurgitate techniques it has learnt before. The advent of AI DJs would be a nightmare for hospitality venues anyway. The premiere of AI technology could inspire audiences to stay sober in hopes of better formed memories, cutting into venues’ drinks profits. Once the concept was no longer a novelty, venues would still be restricted by their own capacity and their customers’ expenditure. If an AI DJ malfunctioned, it might even be quicker to find a human DJ (since there’s millions of them) than to have a dedicated technician at hand. The concept has not proven to be feasible yet, and a substantial amount of faith (and money) would be required to see the AI DJ escape novelty.
To borrow a quote from Melbourne label Butter Sessions, the human DJ is “in control of you losing control”. Relinquishing our power over to the DJ is what thrills so many club-goers. To relegate that power to an AI would remove some of the special yet unspoken human bonds that make clubbing so exciting.