From music sharing to meta DJing
AI will make some creative professions obsolete. DJing will be the first domino to fall. Apps can already mix and curate today, and their level will soon be competitive. What‘s left for DJs?
DJ Rob spins around powerfully, grabs one of the records that had been placed behind him and carries it onto the turntable with one unerring movement. In his shiny silver and gold costume, he looks a bit like a character from a science fiction B-movie, two bright blue eyes flashing behind the visor of his helmet. It‘s a perfectly normal night for Rob, who has been entertaining guests on a cruise ship for TUI for four years. Despite long working hours, he has never complained. That‘s because DJ Rob is a humanoid robot, co-developed by Berlin-based design and multimedia company Pfadfinderei and the robotics team at pi4. In videos and photos, he looks pretty trashy. But Rob is by no means just a gag with which the rather traditional TUI wants to brand itself as a technological leader. It‘s a sign that DJs are already in direct competition with AI for jobs today. [1]
Auto mixing 2.0
While AI production software is getting all the attention in the media, AI DJing is already a reality. Almost unnoticed, Spotify has integrated an auto-DJing feature into most playlists. Simply setting the shuffle function is enough for an AI to now provide seamless transitions between tracks. The result is certainly not comparable to that of turntable magicians like Jeff Mills or Lena Willikens. But it‘s good enough to make the experience a memorable moment. For example, the AI used by Spotify has effective tricks for smoothly bridging trickier passages with echo effects, fading out the bass, etc. [2] It doesn‘t take much imagination to see that this feature will very soon make human involvement obsolete for many smaller events where a good but rather unknown DJ would have been booked in the past.
The wave of digitization in DJing prepared the ground for this initially shocking realization. The sync button has largely devalued the virtuoso element in mixing. The transformation of the DJ into a pop star has meant that fully prepared or minimally live mixed performances are widely accepted. It may sound a bit heretical, but for many of the industry‘s biggest events, the step from a traditional DJ set to an AI mix isn‘t a big deal anymore.
Admittedly, some skills of a human DJ are still beyond the reach of machines today. Most notably, these include those who are critical to the underground club scene that continues to be the very basis of the culture. Above all, this includes the legendary “reading of a room”, which happens on such a subconscious level that not a single DJ has been able to explain it to me in conversation. Currently, AIs are very crudely approaching this ideal by evaluating the reaction to a track by means of the exits and entrances on the dancefloor. Of course, this is not enough and often leads to insane confusion in experiments.
AI curation
Also in the field of curation, i.e. the selection of the titles to be played and the next track, AI do not come close to human DJs. But they are catching up here at a breathtaking pace. Using mathematical analyzes of gigantic mountains of data, collected from thousands of sets over many years, neural networks form a statistical understanding of how DJs think and act.
Of course, this understanding is not dynamic and is therefore never sufficient in practical use. Nevertheless, the collected knowledge is quite interesting even for experienced turntable magicians - is there perhaps actually a formula for good DJing?
Precisely because this know-how seems like a kind of secret knowledge, the first practical applications can be expected here. Apps can make suggestions, question decisions and suggest completely radical new directions or even retain the style of a set as much as possible, similar to cruise control in a car. The „AI DJ Project“has already tried this very thing out in front of an audience via some human-AI back-to-back performances. Their goal is to make the emerging technological change less threatening and make it seem more like a creative enrichment. [3]
Courage for advancement
DJs will have to continue to evolve in order not to become obsolete. They can embark on voyages of discovery with the AI, combining, for example, radical track transitions and phases where the AI is more prominent with classic performances. Alternatively, they may temporarily use software as „autopilots“to integrate additional media into their performances or mix and remix tracks live. Ironically, the increasing use of AI makes virtuosity possible again - namely in the rapid handling of the available potential.
Moisés Horta Valenzuela‘s meta-DJing also goes in this direction. The multimedia artist feeds his software with the sets of the Berlin collective HörBerlin, who have become a kind of meme themselves with their neon-colored, slightly claustrophobic bunker sets. The AI is a further development of AI Jukebox, which reshapes the sets into bizarre, almost creepy underworld constructions - remixes on the level of several complete DJ gigs, so to speak. [4]
Even if some of it seems rather conceptual and cerebral and will neither be used in a scene club, let alone on a cruise ship of TUI, the results show in which direction the journey could go: To new forms and experiences that build on the long tradition, but completely renew it. DJing, in its most original form, is just an evolution of the rather simple sharing of music. Forging deeply moving and innovative experiences out of it is the real task of DJs - and with the help of new technology, they will be able to fulfill this also in the distant future.