“Mind-blowing, ever-changing”: Interactive AI art exhibition HyperCinema goes live in New Zealand
My interview with the CTO of the world-leading AI experience
I had the opportunity to speak with Gareth Hordyk, CTO of HyperCinema, billed as the world’s first (as far as we both can tell!) “live AI experience that places the audience at the center of cinematic narratives”. As soon as we start talking, I know I’m in the presence of someone who’s not just an expert, but a passionate storyteller who’s bringing something to AI.
Gareth has previously created robotic systems at Pinewood Studios, and for Jim Henson’s Creature Shop. It’s exciting to speak with a creative mind that’s fully embraced AI technology (we’re both reluctant to just resort to the cliche of calling it a “tool”; we settle on “collaborator”) while AI is still treated with some suspicion in the industry, and uncertainty by the public.
One of the things he notes is that for many visitors to HyperCinema, it’s their first experience with AI art. Gareth wants to ensure it’s a good one.
Laughter. Delight. That’s how they measure success. Gareth sometimes overhears the hoots of joy and amazement from the audience when he’s working on another floor. It’s also a guiding principle for the team when deciding what directions the AI generations should take. Early on, they realized that the journeys through the AI multiverse that HyperCinema takes participants on — in which they encounter versions of themselves rendered by AI — should take the path of humour, wonder and optimism.
But I’m getting ahead of myself! (Maybe the fractured timelines of Hypercinema are contagious). Multiverse? What is HyperCinema?
HyperCinema is more than an exhibition — it’s a place where art meets science, stories weave in and out of reality, and each visitor becomes the hero of their own extraordinary narrative. It’s a transformative journey.
HyperCinema: Origin story
Based in Auckland, New Zealand (aka Middle-earth — we’re no stranger to overlaying fantastical worlds onto real places in NZ!), HyperCinema is the brainchild of theatre impresario Dr Miles Gregory, and the team at Gladeye.
Miles famously pioneered the Pop-up Globe in NZ and Australia (again, another overlaid landmark). That project brought a simulacrum of the 1614 Shakespearian theatre — and all its magical realms — to Aotearoa (to my international readers, Aotearoa is the contemporary Māori-language name for NZ).
Gladeye, meanwhile, is a creative digital agency in Auckland with a 2023 Emmy Award Nomination in Outstanding Interactive Media for a digital storytelling project published in Rolling Stone. Tarver Graham, CEO of Gladeye, is co-founder and co-creative director of HyperCinema. Gareth Hordyk is co-founder of CTO of HyperCinema, and Operations Manager & Product Consultant at Gladeye. With these creative minds working together—and their strengths in dramatic and digital storytelling—HyperCinema audiences are in for a theatrical, interactive experience unlike any other.
An intersection of AI, art, and the multiverse
But what is HyperCinema? It’s an interactive experience where audiences can explore their ‘alternate selves’ through an AI generated mix of hyper-personalised narratives, told across short films, images, text, and audio.
It’s part Choose-You-Own-Adventure and part benevolent episode of Night Gallery, where all the portraits are of yourself (it’s perhaps telling that the ushers are dressed as futuristic 1920s bellhops from an otherworldly Twilight Zone). It’s a museum exhibit to lives that never happened.
Hitting hyper-speed with AI art
The current season at HyperCinema is called “Enter the Multiverse”. Gareth hints that we’ll see other themes; Halloween came and went too fast, but is a Christmas-verse on the cards? Gareth observes that in the AI space, you have to create with a sense of exhilaration, not look back, and get it out there first (which is how they achieved a world’s first live AI experience).
It’s an approach I strongly agree with; in the ephemeral world of AI art, you need to be first out the gate, create something great, and move on. There’s perhaps a metaphor in how these instant, tailor-made artworks are wiped between each personalized session to protect data: has any gallery before ever deleted and reset its art collection on this scale, or had such singular appeal? It’s both transitory and once in a lifetime. In the past you had to be rich, famous, or aristocratic to have a portrait commissioned; now you can walk through an exhibition of self-reflections, but it disappears behind you.
Nobody else sees these rapidly AI generated masterpieces in which you are the subject (except your “Sidekick”; we’ll discuss the superhero archetypes soon). I can’t decide if that exclusivity — a temporary gallery that’s all about you — is egotistical or humbling, a reminder that “fame is fleeting”. At least there’s a gift shop, should you decide to commemorate the ephemeral experience with merchandise! (You can take photos, but as Eco would doubtless say, souvenirs are the epitome of the hyperreal experience).
Why a multiverse?
In a broader sense, the theme of “multiverse” celebrates the capacity to of AI rapidly iterate infinite variations (the Star Trek philosophy of “infinite diversity in infinite combinations” comes to mind). As any AI artist knows; it’s addictive to keep on generating images. Each one gives us a glimpse of another counterfactual — but often convincingly ‘real’—scene. As of August, over 15 billion AI images have been generated worldwide. It took early photographers 150 years to reach that watershed moment, around 1975.
HyperCinema bundles all our current discussions around identity, reality — and yes, narcissism — into an artistic experience centered around you, but also… not you? Is the AI selfie the height of hyperreality in technologically advanced postmodern society, blurring the lines between fact and fiction?
I don’t know, but it’s one hell of a photo opportunity (taking selfies in an art gallery of impossible AI portraits of yourself would blow Jean Baudrillard’s mind). The multiverse theme is en pointe with the zeitgeist, but also a fun play on the creative potential of AI and the vast array of variables.
Journeys in Hyperreality
HyperCinema takes place across three different acts. First is a trippy immersive projection space, where participants get to see themselves projected four metres high in otherworldly situations. Second is the personalized AI art gallery, and third is a cinema where the participant is treated to short biopics detailing the lives of their parallel selves. The main participant is referred to as the “Hero” throughout the encounter; however you can share the experience with a more camera-shy friend who’s along for the ride as a “Sidekick”. You might bump into other attendees, but they are on their own journeys through the multiverse, passing travellers. But again, I’ve gotten ahead of myself (blame those ever shifting timelines!).
What is a hero’s journey without a physical talisman to usher you out of the ordinary world? And HyperCinema has a doozy: an enigmatic cube. There’s something simple and evocative about this—dare I say it?—cosmic cube. AI and Marvel do keep cropping up together, from Jarvis AI to Secret Invasion:
Is HyperCinema intentionally echoing the tropes of the highest-grossing movie franchise? “Multiverses”, “heroes”, “sidekicks” and “cosmic cubes”? There’s even a multiverse sequence where the participant becomes an astronaut who then de-ages to a baby, in a scene right out of Endgame.
Gareth gives a knowing smile and hints that in a possible future version of HyperCinema, themes could revolve around licensed characters. As my regular readers will know, I’ve been writing that AI driven personalized entertainment is the way forward for juggernauts like Disney/Marvel to meet the challenges of AI head on. HyperCinema gives us a taste of that.
There and Back Again
Accepting the call to adventure, Heroes put their cube into a console and answer a series of personality questions. The results of their digital profile are supposedly stored on the cube, but I suspect that’s smoke and mirrors for AI newbies; it gives them something reassuringly physical: a “totem”.
I like this idea, not just because it blurs the tactile and the AI, but also the sense of embodiment, that their unique identifer is downloaded into the cube. Throughout the experience, heroes use this cube to activate the installations. It’s a wonderful touch, even if it’s potentially flimflam.
The questionnaire is used to guide the narrative, starting with disarming questions like your favourite vegetable and what superpower you’d like to have, building up to more existential questions like what you think is the biggest threat to humanity. I can only imagine the rabbit holes of self-referentiality that answering “deep fakes” might send the algorithm!
In addition to the personality profile, 15 photos are taken; AI experts will recognise this as the rough number needed for a small data set to train AI.
After about 15 minutes, the personalization is complete. The Hero’s visual journey begins in the aforementioned Dreamscapes of Untold Futures (the larger-than-life projections), continues into the Salon of Uncanny Delights (gallery), then culminates in the Cinema of Impossible Histories (cinema).
The Gallery of You
In the art gallery, your cube glows greenly and can be placed to change clusters of framed artworks to your face. Gareth describes participants moving around to trigger all of the portraits— or as many as they can — to reflect them and have a personal gallery (it’s worth noting several Heroes and their Sidekicks can enjoy the experience simultaneously; the custom elements are triggered by your personal movements through the space).
Additionally, each image has an AI text generated description telling the Hero how their alternative self got to that point. Gareth tells me these placards are unexpectedly popular, with people snapping pics of the backstories almost as much as the multiverse images themselves.
Your own “Being John Malkovich” moment
Next, the Hero enters the Cinema of Impossible Histories. Here they are shown a surreal short film featuring a multiverse montage. It’s narrated with an AI voice-over from AI generated text. The sequence is almost a succession of trailers, offering tantalizing sneak-peeks into alternate lives, perhaps including being every member of a family eating dinner (like that scene in Back to the Future Part II, where Michael J. Fox played everyone):
However, you never know quite what collection of biopics you might get: the scenes are personalized and stochastic, so no two are quite the same. Gareth talks about the replayability of HyperCinema, and that one of the surprises was how often they get repeat visitors to relive the journey. It’s wonderfully paradoxical, because it resonates with the ancient Heraclitian notion that “you cannot step into the same river twice.” Each moment of creation is unique, and the software adapts. The work is in constant flux.
As Gareth talks me through the multimedia components, I see so many possibilities that I can’t help but interrupt and ask if they’ve considered evolving HyperCinema into other story-heavy entertainments, like escape rooms with hypertextual narratives. Not yet, Gareth says with a glimmer, and I wonder if I’ve sparked a new direction rippling in their multiverse.
Magic behind the magic: How HyperCinema works
I know my readers will want to know more about the core AI technologies that power HyperCinema, and what methods they might’ve considered (or challenges they faced) when making personalized cinematic experiences.
Personality cubed
First, as an AI afficiando, I’ve spotted something about the cubes. I love the superstitious hold they have on the imagination—this idea that you could download and physically hold your essence, like a hardware high-security wallet for bitcoin—but I’m reminded of the IT Crowd episode where Moss tricks Jen into thinking the internet is kept in a black box with a red light:
I ask whether the cubes really do “contain a personality profile”; Gareth confirms they track the position of the participant. But still, heroes clutch them as if they have a weird, otherworldly power. Gareth notes that when first handed the cube, participants start taking selfies. It’s all part of the show, and I’m reminded there’s a performative, theatrical component.
Are the cubes functional to the generation? Probably not. Are they essential to the experience? Absolutely. Like the bellhops, you buy into a suspension of disbelief when you step through that door. The only question we should ask is the one Alice asks herself in Wonderland (and wouldn’t that be a great theme?): “Who in the world am I? Ah, that’s the great puzzle.”
Picture (im)perfect personalization
As for the workflow behind the whole experience, there’s a patent-pending part of the process, but it’s powered by OpenAI. Gareth is understandably secretive about how the magic works, but it involves a personalized model of Stable Diffusion for each visitor, quickly trained on 15 images. There’s a team of prompt engineers responsible for the 2000+ recipes that lead to the endpoints, as well as some human curation of the results before being displayed. Eventually, they see this evolving to be fully autonomous alignment. However, the occasional glitch or hallucination is left intact, especially if it’s humorous or can be accounted for by the “multiverse”.
For instance, Gareth describes how blue-haired housewives kept cropping up, unprompted, just for one day. Everybody got a 1960s AI Marge Simpson makeover. It’s another point where the multiverse theme ties perfectly into the ad hoc behaviours of AI: highlighting the opaqueness of AI decision-making processes and the fascinating unpredictability that can ensue.
We compare it to Everything Everywhere All at Once, and how conducive the hotdog universe was to priming people to the finger f***ups of AI art.
There’s one more gem I manage to glean for my readers (I feel like Bilbo amongst Smaug’s treasures): the 15 photos that are taken are used to seed an entire fake social media gallery of the last five years — rock climbing, at the beach etc; ordinary things — and this is used as the corpus for training the AI before you see the multiverse. Clearly actual social media accounts aren’t used for privacy and standardisation reasons. I am actually almost more interested in seeing these commonplace, everyday albums that are used to streamline the process; I suspect there’s poignancy to be found in the ordinary (especially as we all missed out on milestones during Covid).
Artistic advice from the multiverse
I also ask Gareth his tips for aspiring AI artists. In addition to the advice to create art rapidly and get your work out there (the name of game with this brand new medium is momentum), Gareth emphasises the importance of using community resources and tutorials, and not being discouraged by comparing your work with the “picture perfect” works that trend online.
“Embrace the errors as part of the creative journey,” he advises. This fosters creativity but also underscores the necessity of adapting swiftly to the AI’s learning patterns and outputs. He also stresses that aspiring artists need to be comfortable with uncertainty and the often unpredictable nature of AI.
More worlds, more options
The final impression I’m left with is one of optimism; it’s a distinctly Kiwi touch that’s been brought to AI art (and makes visiting Auckland worth it, for my international readers, if you’re up for a trip). While I’m sure they’ll take it to the world, HyperCinema is a journey that could only start in NZ.My last question is to ask Gareth what kind of experience or message he hopes participants take away with them. I’m expecting something about making people more comfortable with seeing AI technology as artistic; putting it on the same standing as traditional, gallery-worthy art forms.
“We want people to connect. With themselves, with the possibilities of their own untapped potentials, and with each other, in person. This isn’t just about AI or art; it’s about humans intersecting with technology, sparking conversations, and maybe a bit of self-reflection.”
I can’t help but agree. It’s as much about the in-person social experience as it is about the wonder of AI. It’s a reminder that while technology advances at a blistering pace, it’s the human touch, the Kiwi ingenuity and warmth, that truly makes these ‘authentically artificial’ experiences unforgettable.
For anyone curious about interactive live AI art, or seeking a unique lens to re-envision themselves and their world, HyperCinema is on now at 131 Queen Street, Auckland, NZ with an extended season until 23rd Dec.
Who is Jim the AI Whisperer?
Jim the AI Whisperer offers private coaching on how to write original and compelling content, as well as how to use AI generators to create stunning visuals. If you’re interested in discovering more, feel free to contact me.
I’m also available for podcasts, interviews, fine-tuning AI prompts, and creating prompt libraries and professional AI artistry for companies.
