Vanishing Mediator
The Vanishing Mediator is an interactive robotic installation that observes its environment and transforms sensory input into a generative visual narrative. A four-legged robot equipped with a 360° camera scans the space and its visitors, transmitting data to a custom software system. This system, developed by the artist and named bunraku uses three interconnected AI engines to process the data and generate a continuous stream of text, images, and prompts in real time.
By choreographing this loop of perception and production, the work invites us to consider the nature of machine creativity. Are we witnessing autonomous generation, or merely projections of human desire mirrored back through algorithms? The system appears to create, but it also reflects, both aesthetically and ideologically, the complex dynamics of authorship in computational environments.
The title and conceptual foundation draw from Fredric Jameson’s idea of the Vanishing Mediator, a figure or process that enables structural transformation before fading from view. Extended by Slavoj Žižek into psychoanalytic and political realms, this logic frames mediators as necessary yet ultimately transient. The revolutionary who disappears after the uprising, or Lacan’s objet petit a, which sparks desire only to remain unattainable, are emblematic of this phenomenon.
This theme echoes powerfully in the Shahnameh, the monumental Persian epic by Ferdowsi. Composed around 1000 AD, it recounts the mythical and historical past of the Persian Empire and became a vehicle for cultural preservation during transformation. With its multitude of versions, shaped by the hands of different patrons, the Shahnameh functions as more than a literary artefact; it becomes a shifting mediator of identity, memory, and meaning.
In this context, The Vanishing Mediator poses a central question: What must recede for our desires to be fulfilled? The work reflects on contemporary longings for machines that are not only intelligent but also generative, systems that move with modular precision and recognise patterns with an almost non-linear generative responsiveness.
Technically, the installation brings together Boston Dynamics’ Spot robot, Microsoft Azure’s cognitive services, OpenAI’s language model, and Stable Diffusion’s image generation, all coordinated in real time using vvvv gamma. The result is an integrated robotic system that does not simply perform but responds, creating content that shifts as it perceives.
Ultimately, the installation acts as both autonomous agent and mediated mirror. It produces a mythic feedback loop: a robot retelling fragments of a world it observes, one that is already shaped by our presence, our data, and our desire to be seen.
—
Supports:
Creative Robotics – https://creativerobotics.at/
FestivalX – https://festivalx.ae/
DEWA – https://www.dewa.gov.ae/en/
Qapture – https://www.qapture.at/
—
Tgas:
QuadrupedRobots, GenerativeAI, BehaviorTree, CreativeRobotics, psychoanalysis, OpenAI, Transference
—-
Software:
vvvv gamma – https://visualprogramming.net/
Bunraku – https://bunrakuproject.xyz/
—
Hardware:
Boston Dynamics Spot
INSTA 360 X3 Actioncam
Custom build micro controllers
Windows Tablet
PC Station
LED screen
