How can we know what to expect when AI can so easily sell us a dream?
Image: 'Open Beta', created and performed by Chivaree Circus. Photo by Dom Martin Photography.
At the end of February 2024, a story broke about a family Wonka experience gone wrong. Within hours, global news sites were reporting on a starkly desolate warehouse in Glasgow sold as an immersive event. The camera phone images were bleak. There was a schadenfreude of watching the actors dutifully try their best to temper this chocolate disaster while the attending children were given a lone jelly bean. Just like with Fyre festival, a big propellent to this story gaining so much attention was the yawning gulf between what had been promised in marketing content released to sell tickets and the reality of a very poorly planned event from people who clearly didn't know what they were doing. While Fyre had employed a boat load of models to push people to their island paradise, it quickly transpired that House of Illuminati (the company behind this ill-fated adaption of Roald Dahl's classic) had simply used AI to generate advance images. All of the ingredients were present for a very 2024 kind of viral story.
Now that the dust has settled, I can see exactly why families bought into the idea of this event. The pictures used to sell it are incredibly exciting. They are beautiful, colourful dreamlands. You can imagine getting lost in candy cane forests as a swirling chocolate river bubbles beside you. Who wouldn't want to bring their children to something so magical? Of course in hindsight they are too good to be true, but most of us still expect to be treated honestly by those selling us experiences and we are not used to mistrusting every piece of digital content as a first reaction. We know AI tools are out there to create and manipulate what we see online – even the Royal Family appear to be trying their hand at it – so we probably need to start becoming more discerning. Right now AI is coming to shape our world and for some in events, nightlife and the wider artistic world it has very clearly already arrived.
Image: 'Open Beta', created and performed by Chivaree Circus. Photo by Dom Martin Photography.
I spoke to a range of creatives with different experiences. Some were very positive with one event organiser confidently exclaiming: “people are getting pushed out this idea from the media that AI is going to take your job and that's not true.” There are certainly some very clear benefits. Vicci Moore is now using AI to translate her dance class videos to other languages which opens up a huge new market for her online school. Her brother Sam, uses generated images for pitch decks for event clients saving masses of time that would have been spent searching for relevant references otherwise. However, Emilie Woodrow told me that while it can be helpful to get a “general overview of what a client is thinking, it can produce some wildly unrealistic ideas of what can actually be built.” For her, as long as everyone understands that then it’s not a problem. But others are more frustrated. Janna Rogers who runs Planet Gold Decor is “so sick of AI 'art'. We are drowning in endless, impossible images. People are getting ridiculous ideas of what might be achievable from creative suppliers.”
One digital artist told me she was “essentially on the verge of giving up as there was no longer a space to create work when AI could do it instantly for free with vast levels of details and texture that could not realistically be achieved by a human artist.” This was echoed by Dean Samed who had a profitable boutique stock photography website with an asset library. He describes how “AI came and ripped through the entire industry within minutes and online platforms like Adobe Stock and Shutterstock were flooded overnight with AI generated content.” He sold his business at a loss “to get out as fast as humanly possible.” With many years of work, love and labour, paying creatives, freelancers, makeup artists, models and studios over for him. He was angry that people could now generate assets based off of his IP for free, saying “he actively despises AI” and that he feels there is no longer any joy in what for decades was his main passion in life. Constanza Ruff – a voice over artist also expressed concern that: “Companies are using people's voices and slightly tweaking them so they can't be sued for theft. Equity has a template for clients, but some refuse to use it. Half the books on Apple now are narrated by AI and of course, they are cheaper or free to download. This could be the end of voice acting as you can basically create anything you want with AI.” She points to Stephen Fry's voice being “stolen” from the Harry Potter Audio books and used for a documentary without his knowledge last year. “He's one of the most recognised voices in the planet. “ She said. “If it can happen to him what hope do the rest of us have?”
In January, I produced a new circus show OPEN BETA with my company – Chivaree – that used AI entirely to generate script, choreography, music, costumes, visuals, marketing and ideas at every level of the production. We were wholly upfront about this with our audience, asking them to experience the cognitive dissonance of not always knowing what is human inspired and machine generated. We were awarded funds from the arts council to research this. Crucially we paid all of the creative team as we would on any other project to work alongside AI and use it as an experimental tool, instead of simply replacing them to save on costs. The show sold out and was a success with audiences curious to know what could be achieved with the current tools. However, some online commentators were deeply skeptical, asking if we were selling out to the machines and reducing the possibility for human creativity. I understand this view, but AI is here whether we like it or not, and rather than shying away from something that is actively shaping our world, we chose to creatively critique it and try and understand it's potential.
A few other shows are exploring how to use AI in live performance. Notably, Company Wayne McGregor presented version 95 and 96 of their show Autobiography at Sadler's Wells (12+13th March 2024). Using a new AI tool developed with Google Arts and Culture trained on hundreds of hours of McGregor’s choreographic archive. By integrating AI into their work, performances are continually evolving and never quite the same.
From doing OPEN BETA, I can say confidently that generative AI is more than a flashy toy or cheap trick. If you are creating content for the digital world like images, speech, marketing copy or music – it is already impacting the viability of creatives to continue earning a living. Several people working in the film industry also expressed serious concerns that they too would soon be under threat. If however, you are producing experiences that come into the physical world, for now at least, you still need the human expertise to deliver them. AI can give you a description of choreography, but it may be physically impossible to realise the movements. It can give you a script, but without a decent writer or director to edit and interpret, you can quickly end up like the Wonka experience with a preposterous sub-plot about a rival chocolatier who lives in the walls. More common though is just something artistically uninteresting and mundane. Large language models are very good at predicting the most likely response which by definition means generally they won't produce a fresh or new take.
For now if you see things online it's time to start approaching with a degree of scepticism. If you are looking at an image, are there too many repeating patterns? Are there inconsistencies in the resolution? Consider the context and plausibility of what you are seeing and if you are buying a ticket to event, look at the companies track record of delivering them!
Article written by Edward Saling, co-founder, Chivaree Circus.
chivaree.co.uk
Image: 'Open Beta', created and performed by Chivaree Circus. Photo by Dom Martin Photography.