Using AI to transform DJ set visuals

CR speaks to Boiler Room art director Joshua Wiley and motion designer and artist Hayden Martin about their ongoing collaboration that uses neural networks and Boiler Room’s extensive image archive

When Boiler Room started in 2010, founder Blaise Belville set up his decks in a 1930s-style boiler room, taped a webcam to the wall and began streaming his sets online. The weekly sets grew, and now 13 years later, Boiler Room has put on over 8,000 performances by over 5,000 artists across 200 cities.

A counter to the mainstream, media-driven music at the time, the raw, experimental approach of Boiler Room still remains and as such the brand has created a recognisable aesthetic, originating in its low-fi approach. This style has become so intrinsic to Boiler Room that recently it began using its own footage as visuals to accompany DJ sets and live events.

“We’ve always had visuals at our events. When Boiler Room first started we mainly used found footage from the world of electronic music,” says Joshua Wiley, art director at Boiler Room. “But at this point, over a decade after Boiler Room started, if you look for a DJ set online, you’re likely going to find Boiler Room.”