Medium: GPT4o, Node.js, WebSpeechAPI
A.R.T. (Artificial Representation Technology) is an interactive conversational AI system designed as a synthetic extension of my artistic persona.
Trained on my personal archive of artworks, research papers, and published writings, the system engages visitors in real-time spoken dialogue using a cloned voice model while dynamically pulling matching visual documentation as the conversation unfolds. Rather than acting as a passive portfolio, it transforms the traditional artist archive into an active, asynchronous dialogue between creator and audience.
The installation debuted in Singapore during the launch exhibition for Vector: Drop 001—a curated physical and digital publication by Metalabel that champions new paradigms of collective creative release and contemporary digital art. Hosted at Mantala Club, the launch gathered international artists, technologists, and collectors to explore physical activations of forward-thinking digital works.
It was designed as a physical kiosk, the installation features a white pedestal equipped with a gooseneck microphone and interactive press-to-talk button, inviting gallery visitors to speak directly with the AI proxy. A dual-display setup anchors the physical interaction: a vertical portrait screen renders the live chat transcript and synthetic persona interface, while an overhead monitor dynamically projects high-resolution artwork images, schematics, and video documentation pulled in real time as specific topics are discussed.
At its core, A.R.T. investigates delegated representation and the emerging boundary between human creator and synthetic proxy. Most portfolio sites treat an artist’s history as static documentation, but this project tests how effectively machine intelligence can convey nuanced creative intent without flattening the human origin. By synchronizing low-latency voice synthesis with live visual indexing, the installation creates a more intimate, reflective dialogue between the work and the viewer.
The artwork was powered by GPT-4o and custom knowledge orchestration engines, the platform integrates real-time speech recognition via the WebSpeechAPI with high-fidelity voice synthesis and Node.js context tracking. During the exhibition the artwork was live under the following URL artificialrepresentation.tech.