Light / Dark mode

Ghostwriter: Communicating with Jo Ho’s a developing slate

Wires spill onto the gallery floor, immediately suggesting that the interaction I’m about to have with this work by Singaporean new media artist Jo Ho will be a tech-driven one. Ho’s work often explores “digital corporeality” — referring to the interconnected state of the digital and physical worlds — and the role of agency in an era of rapidly advancing computation. 

Still, the setup of this work surprises me. I approach the plinth with a mix of curiosity and some vague recognition. Though I have not encountered anything quite like a developing slate (exhibited at the gallery Ames Yavuz as part of the Paratext show curated by Clara Che Wei Peh), something about it feels strangely familiar. Where have I seen this before? As I search my memory, it hits me that the setup resembles archival scanner systems used for digitising books.

Observing the web camera, tablet screens, aluminium supports, acrylic spacers, and what seems to be a piece of paper, I instinctively lean in to take a closer look. The two displays, resembling the open pages of a book, draw me in, so much so that I have to resist the urge to reach out and touch them. On phones, tablets, laptops, billboards, and even public transport directories, screens are now everywhere. But how often do we stop to think about our relationship with these interfaces, and how the things we read, scroll, type, and send traverse digital space? 

Installation view of a developing slate (2024), acrylic sheets, digital screens, webcam, ink, and wax-resin composite, dimensions variable. All images courtesy of the artist and Ames Yavuz, with photography by Ken Cheong.

As I stare at the two opposing screens, I recall a conversation with a friend about how technology only feels truly functional to the wider public if it’s like magic — effortless, uninterrupted, and real-time. Ho’s piece embodies a sense of this “tech-magic,” with its book-like familiarity conjuring memories of leafing through pages and immersing myself in other worlds.

But just as I am about to dive into the text, I hit a barrier. What I see on screen doesn’t seem to refer to any alphabet or logographic characters I have encountered before. I peer into the work again, trying to decipher the ethereal palimpsest of symbols and lines that fade in and out of view. Like a futuristic, alchemical tome, the work seems to serve as a medium to another realm of communication, holding something just beyond immediate comprehension.

Ho explains that the digital renderings I am trying to decipher come from her handwritten notes, which she fed into an AI-trained text-to-image and video editor called Runway ML. The platform transforms her notes into spectral renderings of what machine learning — adaptive algorithms that learn from data patterns — predicts an annotated page should look like. Each pixel builds upon the machine’s trained model, extending the image based on probable continuations, or what the algorithm perceives as a sensible extension of what came before. Stringing pixels the way humans string symbols, sounds, and gestures to communicate.

After printing and encasing this rendering in a wax-resin composite, Ho scrubs away the paper pulp, leaving only a translucent imprint, which will be scanned to become another machine-generated assumption. This time, her prompt asks the machine to render a video of a page being flipped. What the computer manifests is a shadow, cast by nothing, gliding across the screen. A small web camera live-streams this ghostly flipping, adjusting the line thickness on the second notebook display where the “phantom hand” interacts. It feels eerie, as if I am watching something being archived while it is still in the process of becoming.

In the back-and-forth between the physical and digital, manual and machine labour, what began as a legible note gradually morphs into an illegible language. Undeterred, I keep searching for meaning in what I am “reading.” Finding meaning is central to human cognition, as we are hardwired to make sense of the world around us. In Ho’s work, AI does not aim to produce outputs that we can understand; instead, it highlights the complexity of comprehension itself. Moving between sense and nonsense, in a cycle of decoding and recoding, the machine does not need to understand the input — it processes the data, stripping it of its human-centred significance, and reconfigures it based on probability — drawing its algorithmic response from thousands of predecessor texts.

Exhibition view of Paratext at Ames Yavuz.

Ho’s work surfaces this tension between different systems of comprehension, challenging our understanding of communication by juxtaposing ancient and contemporary forms of information transmission. It raises questions about how we, as humans, perceive information, and how a non-human system like AI interprets it, acting as translator of a language it doesn’t speak but has been trained to process. 

a developing slate prompts me to reconsider how meaning is made and translated, extending beyond human networks. As I write this, a natural language processing chatbot is checking my text for grammar and flow — a meta moment of AI writing about AI. Since researching Ho’s work, the concept of trans-human sense-making has been lingering in my mind. If inter-human communication already results in so much miscommunication, how perplexingly complex it must be for an emerging system that merely simulates our communication patterns.

___________________________________

Paratext was exhibited at Ames Yavuz from 20 July to 17 August 2024. Find out more about the artist at j-o-h-o.com

Support our work on Patreon