Experimance

Experimance is ongoing series of experiments, the most recent of which is on display at the Factory Media Centre in Hamilton from Aug 7- Sept 29, 2025 (noon-5pm M-F).

You can also see prints from the series at the Tarragon Theatre in Toronto all summer, hours are 11am-5pm. (Don’t forget to check the staircase!)

Visitors shape the white sand with their hands. A depth camera reads the topography, and in response, AI-generated satellite-like images are projected onto the surface. These visuals shift to reflect changes in the landscape, blending natural forms with human infrastructure, circuitry, and computation. As you sculpt, the installation speaks. An AI voice agent engages visitors in real-time, inviting them into conversation about climate change, artificial intelligence, and the future we’re creating—whether we mean to or not.

–

Artistic Statement

I am filled with hope, dread, and guilt about the human experimentation on our world and the accelerating pace of technological change that is our children’s inheritance. This recklessness now extends to AI, as we rush toward machine intelligence, ushering in new species, tools, and ways of thinking. It reminds me of flying over a vast city—marveling at what we have built, while grappling with its bloody price: our past, present and future sacrifices. My work is about knowing and the awe and horror of that knowing.

Experimance is a series of experiments that help me think more deeply about the nascent rise of machine intelligence and the recurring historical disasters that arise when humans act without consideration for the consequences—following the money, not the richness—the most catastrophic being anthropogenic climate change. The struggle for climate justice is a slow sketch of our current disorganized corporate-led stumble into the creation of alien machine intelligence, but both leave us with the despair of relentless repeated mistakes, of waiting too long and not long enough.

Experimance frames this thoughtlessness visually as the impact of humans on their environment, but the imagery contains allusions to computation structures that (em)power our society and our current and future machine tools/children.

I chose satellite viewpoints, not just for their soaring perspective, but because they give a feel of omniscience, of power, like the computer games we play and the spy satellites that watch us. Yet it fundamentally flattens, ignoring depth, and thus the illusion of being all-knowing is truly blinding.

My own “experimance” involves daily use of AI tools to explore and see how they change me. AI code assistants help write the majority of the code, and AI is used to speak to the audience, generate some of the environmental sounds, and all the images in response to the topology of the sand. These images remix and collide the things we need to live with those which we need to compute, reflecting the growing impact of human development on the world.

It is a tragedy that the arts community has recently begun to reject AI art-making tools, and I believe that it is incumbent on artists to explore deeply with open hearts, to better understand how they change and imperil us. With thoughtful engagement we can continue the artistic tradition of inspiring a future into becoming.

How it Works

Experimance is powered by a suite of real-time technologies, orchestrated by custom software developed over two years:

Depth & Vision Sensors
A Realsense D415 depth camera monitors the sand’s shape, and a webcam detects visitor presence.
Image Generation
The system uses a Stable Diffusion XL AI model (Juggernaut XL Lightning), and two LoRAs (Drone Photography and a custom built one for the Experimance style) to generate visuals in real time, based on the detected sand shape and a set of curated “biomes” and eras of human development.
Sound & Music
Ambient sounds, many of them AI generated, match the visuals, while original music by Daria Morgacheva aka Garden of Magic underscores each era’s emotional tone.
AI Voice Agent
Using GPT-4o for dialogue, AssemblyAI for text-to-speech and Cartesia AI for voice synthesis, the installation converses with visitors. It answers questions, reflects on its own nature, and guides interactions. The dialogue structure is tightly designed to encourage curiosity and respect silence.
Software & Control
The installation runs on custom Python software integrating all elements—vision, audio, sensors, and generative models—written with the assistance of AI coding tools (Claude and o3). SuperCollider handles audio and music. Pipecat manages conversational state and function calls.

Credits

Concept, Design & Software: Ryan Kelln
Music: Daria Morgacheva aka Garden of Magic
Technical Support & Photography: Benjamin Lappalainen
Curator: Gladys Lou
Text-to-Image AI: Stable Diffusion XL base, Juggernaut XI Lightning custom model, Drone Photography for XL LoRA by LordJia, custom Experimance LoRA by Ryan Kelln
Voice Agent & Architecture: Pipecat, OpenAI GPT-4o, AssemblyAI STT, Cartesia AI TTS
Audio software: SuperCollider
Sand: Ethically sourced marine Aragonite from The Bahamas