New technologies have been displacing humans in the workplace since the industrial revolution. Artisans have seen their crafts taken over by soulless machines that work longer, faster, and arguably, better than skilled craftsmen. Profiteers have encouraged these advances while the hardworking men and women have often voiced their opposition vocally and violently.
Historically, the Luddites of the 19th-century English textile industry opposed the use of certain types of automated machinery, and often destroyed the machines by burning them down. They protested manufacturers who used machines in “a fraudulent and deceitful manner” to replace the skilled labor of workers and drive down wages. In the modern entertainment industry, our new concern is that of Artificial Intelligence. We have long thought that creativity was the sole creation of the human mind, but as we outsource more of our tasks to our silicon-based counterparts, we are learning that the complex algorithms of human minds, can be recreated by microchips as well. When AlphaZero mastered chess in four hours, we all had to step back and take notice. Now it’s time for us to discuss if AI has a place in lighting design and operation.
The pAIns
We can all accept that AI will have the ability to replace our jobs sooner or later. We have already seen incremental changes with timecoded shows, automated motors, and even auto-tracking follow spots. These have replaced some skilled operators with software and algorithms. As my colleague, Jason Giaffo, of Giaffo Designs, pointed out online, “When we replace the need for lighting directors on the small level, we kill opportunities for them to break into the industry to reach the higher levels.” He poses a very important question, “Is the financial benefit provided to lower end bars, where shows are fine without automated lighting worth the tradeoff of opportunities lost to aspiring professionals?” “This isn’t about snuffing new technology for self-preservation,” Giaffo says. “I’m a coding nerd myself – I live for pushing these envelopes. This is about having critical conversations about how AI is going to be allowed into what we do. What will constitute fair game for training data? What are our rights over that data as creators? How many expressive opportunities will go extinct to “good enough” cranked out of AI, and how fast? What does it mean to utilize AI to aid in artistic expression rather than starve that expression? We need to start and sustain these conversations across our industry, before we are locked eyes with bigger issues, not in retrospect. AI isn’t ‘good’ or ‘bad’. AI is powerful. It needs to be handled as such.”
These types of questions have been asked many times throughout history. Often what looks like progress, has a companion element of destruction. For every gain we make in convenience, whether it’s AI or autotune, we need to ask what we are consequently losing in expression.
The gAIns
To answer Giaffo’s questions, I reached out to Justin Love, Co-Founder and CEO of Limbic Media. Limbic Media has been innovating in the commercial lighting space for the past five years. They released their first consumer product, an intelligent lighting controller called MaestroDMX, in the Fall of 2023. MaestroDMX is an “autonomous lighting designer-in-a-box that listens to music and makes decisions like a professional lighting designer.” MaestroDMX was shown at LDI’23 and won Rave Pubs Best of LDI, “Best Product for DJs.” Love says that their team has graduate-level education in Digital Signal Processing, Music Information Retrieval, and Metacreation, a field focused on endowing machines with creative behavior. In addition, most of their team members are also artists with active and diverse practices in music, light art, traditional visual arts, design, media art, and interactive art.
Love says, “A couple of years ago, we started thinking of ways to bring this innovative technology to more people via a consumer product, and the idea for MaestroDMX was born. A successful Kickstarter validated the product market fit, and we spent the next sixteen months developing, manufacturing, and launching MaestroDMX.” Love continues “MaestroDMX was designed for individuals and situations that would benefit from improved lighting design but don’t have the resources for professional support, including wedding DJs, small club DJs, hobby bands, small music venues, karaoke bars, spin gyms, music streamers, and more.”
MaestroDMX should not be confused with sound-activated lights. Sound activated fixtures don’t talk to each other to create a cohesive lighting design. MaestroDMX uses more advanced audio analysis algorithms that examine a broad set of audio features along with an advanced decision-making meta layer and generative lighting engine, enabling it to create more sophisticated lighting designs. MaestroDMX is also designed to understand how music is structured, is versed in lighting design fundamentals, and emulates the decision-making process of a real-life lighting designer.
Love wanted to make it clear that MaestroDMX is not ripping off the works of current designers, rather, “Without giving out too many trade secrets, none of our training data comes from lighting designer show files, stems, or other data sources produced by LDs, e.g., videos of light shows. Our audio analysis engine utilizes machine learning. The data for this came from academic databases explicitly created for this purpose. We plan to explore more AI-based innovations in the future.”
When asked if he would consider using AI software to work alongside human creativity rather than replacing it long-term, Love replied “I think this will always be the case; even with MaestoDMX, the user needs to think about what kind of light fixtures and arrangement they want to use and what kind of light show they want to coordinate with their music, e.g., color palettes and general behavior.” He made sure to emphasize that “MaestroDMX also has more advanced features for users who want more creative control, e.g., setting different sets of lights to different fixture groups. Users utilizing these features will produce more sophisticated and unique light shows.”
Some common ground.
The existence of MaestroDMX does raise concerns about replacing human lighting designers and operators. Love, Giaffo, and I agree that this is a valid concern that transcends all industries and occupations, including our own. Love says, “These concerns should be taken seriously as a society, and we need to have open and respectful dialogues about the potential impacts of AI-based technology.” Even Love is open to putting safeguards where necessary, including the possibility of industry regulation.
I must agree with Giaffo when he says that this technology deserves the utmost scrutiny and extensive discussions. Displacing humans for technological convenience has had mixed outcomes in the past. I also agree with Love when he says that this technology is a long way from replacing human artistry on a large scale. Personally, I have been to many venues that have opted for a static look or a pre-programmed rainbow chase instead of putting any effort into making their lighting appealing or interactive. Several years ago, I even imagined a product like MaestroDMX could take over while I went for a pee break during long nights.
Hopefully, we can agree that technology will never diminish the human impulse to imagine and create art. The point of this discussion is to make sure that technology is providing us more tools to make art better and not to destroy entry-level positions.
Also by Chris Lose: