sent(i)ence is an immersive performance-game in which two audience members interact with an absurd assemblage of poker, tarot, and AI to explore themes of self-identity, authorship, and agency within language.
Participants use a limited 52-card deck of phrases to write an ostensibly personal statement of self-reflection, which is then "autocompleted" by an AI that completes the sentence in a confident tone. Before being allowed to receive the response, however, participants are required to prove themselves as human through a series of increasingly absurd and existential tests.
The piece examines how we as subjective selves bring our own meaning to words—in particular, the word "I". When an artificial intelligence uses the word "I", what does it mean? Is it the same word as when I use it?
How do we as humans search for meaning in a world where language has become a post-scarcity resource?
sent(i)ence is my master's graduation piece and was performed at the V&A museum as part of their Digital Design Weekend 2025; a followup to my piece shown there the previous year, Conversation(A/I)symmetry. It is the culmination of my Masters in Computational Arts at Goldsmiths University, and continues to refine my practice of using familiar games and settings to explore language and agency in the age of AI. This time I brought in my filmmaking background to construct a set, script, and performance. As performer I embody the blurred, overlapping roles of poker dealer, tarot reader, AI avatar, bureaucrat, and artist.
I tried to construct a shared moment with my audience, one that would pierce through the deluge of AI-driven technobureacracy and infinitely-generated wisdomslop, to allow me to share a moment of humanity with them. Thank you to everybody who shared their moments with me.
Two players use their mouths to swipe on words suggested by an AI, building a first-date conversation one word at a time.
The conversation is streamed live to a screen, for all to see, while other AIs write judgemental social media comments about how the date is going.
Conversation(A/I)symmetry is an installation game I made, powered by ChatGPT. It is a satirical exploration of the increasing reliance on AI to express ourselves, and the ways in which technology mediates human connection.
The game took place at the V&A Digital Design Weekend on September 20-22, 2024. @vamuseum
#chatgpt #ai #computationalart #artgame #installation #installationart
sent(i)ence originated with a question: how closely related are the ability to use language and the awareness of a self? When ChatGPT says “I,” who is it referring to?
Technically, the work builds on the use of sparse auto-encoders to identify and influence individual neurons in a large language model. When the gemma-2-9b-it model reads the word 'love', for example, neuron 10,391 on its layer 31 becomes highly activated. (This neuron is labelled “expressions of love and affection” on neuronpedia.org) By multiplying that neuron’s weight, the model becomes more likely to mention 'love' in its responses, even to irrelevant prompts. This process is known as steering.
I brought this idea into a hybrid game of tarot and poker, where each of the major arcana is linked to the neurons that activated when the AI read through AE Waite's canonical interpretation of that card. For my deck, I filled the designs not with imagery but text: those sentences in the AI dataset that most strongly activate those specific neurons. Real sentences, once written by real people, now compressed into statistical features it has learned to recognise and reproduce.
Meanwhile, participants build first-person sentences by playing short, ostensibly personal phrases written on a deck of poker cards. An AI, steered by drawing a tarot card, then takes up the decentralised "I" of the sentence and "autocompletes" it in a more confident tone.
Somewhere deep in the AI model lies layer 31, neuron 98,898: “the pronoun ‘I’ to identify personal perspectives or experiences.” It flickers each time the AI encounters the word "I" - but there's something unique about this word. As the AI scraped through its petabytes of data - absorbing countless sentences, each written by a real person - each of those "I"s was a point at which language transcended itself: an instance of a consciousness using language to refer not to something external, but to itself. A sentience represented by an i in a sentence. This is what Hofstadter called a 'strange loop,' and while the AI can recognise and reproduce that linguistic pattern with perfect fidelity, what it cannot do is mean it.
"Something Went Wrong" is a short series of photographs I took on a recent hike in Taiwan, superimposed with recent errors I've encountered while signing in, completing two-factor authentication, and proving my humanity.
Here’s your Instagram reel description copy, written in an effusive, enthusiastic tone that spotlights the tech:
🚨 One click. Infinite possibilities. 🚨
This is what happens when you paste a poem into an AI-powered video generator.
I used Brion Gysin’s “I Think Therefore I Am” — and in seconds, it spun out a full corporate presentation, titled "Exploring Descartes: The Philosophy of 'I Think, Therefore I Am.'"
No prompts. No edits. Just raw script in, fully-formed video out.
The AI didn’t just make a video — it understood the reference, captured the tone, and structured a whole philosophical breakdown.
We’re living in a new era of knowledge production — where complex ideas can be visualised, shared, and scaled at the push of a button.
This is more than just a tool — it’s an accelerator for human thought.
📽️ Script in. Meaning out.
#AItools #BrionGysin #Descartes #generativeAI #automatedvideo #philosophyandtech #creativetech #futureofeducation #languagegames #computationalart #GPTpowered
Let me know if you’d like to tweak the tone, shorten it for TikTok crossposting, or add a call-to-action like “DM me if you want to try it.”
That’s a wrap on Digital Design Weekend 2025 ✨
Over three days the V&A welcomed immersive installations, screenings, workshops, and interactive installations exploring how we can reimagine digital cultures through design, ecology, and immersive arts.
Huge thanks to all the incredible artists, designers, researchers and technologists who shared their work:
· Michael Akuagwu · Julie Freeman · Tamaris Ellins & Fashion Innovation Agency at London College of Fashion · Xach Hill · Joseph Wilk · Ciara Neufeldt · Phoebe Hui · Joseph Lindley & Roger Whitham · Violet Owens · Thomas Fox · V Buckenham · Beatrice Asia Taylor Searle (Peckam Digital) · James walker (London Pattern Club) · Jessica Starns · Studio Above&Below · The Lumen Prize
We also showcased works from student cohorts from the RCA’s MA Information Experience Design (Gayle Chong Kwan, Fatimah Nader Alabed, Audrey Renouf, Xianliang Ye), Goldsmiths Computational Arts (Marcus Round, MDamager, Qiang Hei and Mikhail Aaron), and Camberwell College of Arts (Melissa Li, Nikos Antonio Kourous Vázquez, Eryn English-Polch, Lyra Robinson).
The displays by Michael Akuagwu and Studio Above & Below remain live until mid-October.
Thank you to all V&A colleagues who supported the event (special thanks to Devanshi Rungta and Nick Murray) and to everyone who joined us - artists, partners, and audiences - for shaping this year’s festival and the conversations around art, design and technology.
“It is unfair to claim that an elephant has no intelligence worth studying, just because it does not play chess,” said Rodney Brooks, a roboticist and CTO behind the Roomba. He was making a metaphor about artificial intelligence in his 1990 paper, titled "Elephants Don’t Play Chess".
In this piece of mine, titled "大象下棋 (An Elephant Playing Chess)", an autonomous robot is styled after the Elephant (象) piece in Chinese chess (象棋) and programmed to follow the rules of that piece.
It is then placed into the middle of a context for which its programming is out of place – a game of chess (国际象棋).
Visitors are invited to play a game of chess with each other, but must accommodate the disruptions brought about by this unexpected guest.
What does it mean when the elephant inadvertently pushes one piece onto another, or when a piece’s position becomes ambiguous? Each pair of players is invited to decide this for themselves.
The work examines games as a measure of intelligence, and as a form of language. It looks at the miscommunications that can arise across cultural barriers, and asks how we as humans navigate them. It continues my practice of using games to stage interactions between AI and human agency.
Brooks also offered his own take on the "three laws of robotics". His second law states: “When robots and people coexist in the same spaces, the robots must not take away from people’s agency, particularly when the robots are failing, as inevitably they will.”
#robotics #chess #computationalart #arduino #robots
Two players are invited to play a game of Go (Weiqi). A camera is trained on the game, and a machine learning model extracts the game state. This data is fed to a diffusion model, where the structure of the stones guides the composition of dreamlike imagery that is projected live alongside the players.
In 2016, AlphaGo finally beat the world's best Go players, and our new age of rapidly-exploding AI had begun. AlphaGo used human-style intuition that could at last be called 'creative', as opposed to the brute-force algorithms that excelled at Chess and which dominated 20th century computing. The new 21st century will be defined by creative, generalist AI, ushered in by AlphaGo. The wave of image-generation AI soon followed, and this project attempts to explore that lineage.
The project continues my fascination with games and the artefacts they generate, and continues using AI to ponder the workings of human intelligence and neuropsychology. It takes as an aesthetic starting point the idea of the "Tetris effect," a phenomenon where intense and prolonged immersion in an activity (such as playing Tetris, or indeed Go) temporarily reshapes a person's visual perceptions, even in contexts unrelated to the activity itself.
The idea of this effect invading our dreams is very evocative to me, especially paired with the fluid, dreamlike creations of generative AI. Like a Go player sleeping in between days of a tournament, I imagined AlphaGo sitting in its dark server farm, playing Go against itself and hallucinating imagery - a creation myth prefiguring its image-generating descendants like Stable Diffusion.
Making imagery from Go games...
I use the image2sgf python library, modified to read from a live webcam feed of a 9x9 go board.
This then feeds the game state over OSC into TouchDesigner, where I have StreamDiffusion generating imagery, using ControlNet to anchor the imagery to the state of the game in front of it.
I use various parameters in TouchDesigner to transition smoothly between different prompts, seeds, and weights over time.
#touchdesigner #stablediffusion #ai #go #baduk #weiqi
I am a strange loop. This is not an apple.
An afternoon sketch inspired by the Jetson Orin NanoOWL starter project, and a meme where someone fed a Magritte painting to an image recognition AI.
In "I Am A Strange Loop", Douglas Hofstadter uses video feedback (a camera pointed at its own screen) to illustrate his theory that consciousness arises from cascading loops.
The mirror test is a way to assess consciousness. Can an algorithm recognise itself?
There is a theory that consciousness arises from language, not the other way around. As an entity forms the ability to label other entities as "man" and "apple", and arrange these words into sentences, syntactic space opens up for an entity that must be labelled "I", and from this the first sparks of consciousness emerge.
I remember the first time an entity labelled me with "man". I was only 17, just a boy, but he couldn't see my face, and my tall height probably made me seem older. "Let this man through," he said to his child. I'm now much older and still don't feel entirely comfortable labelling myself with "man".
Nothing you see here is a man, or an apple, or a mirror. It is only pixels.
A reaction to the way that many approach ChatGPT as a way to seek guidance from an omniscient, ethereal consciousness -- a pursuit that has been ongoing for centuries.
#chatgpt #ai #arduino #ouija
My travel video from Beijing and Inner Mongolia!
By coincidence I had three dear friends all visiting their home towns in Neimenggu at the same time, so I decided to visit them all, as well as my Mandarin teacher in Beijing.
With gratitude:
@bbbijinggg@vio_la_liang@luo6342@oilyoo_w
Music by @chineseamericanbear
#chinatravel #travel #beijing #innermongolia
