ANDROID PLAZA - Cavalier I
Postpainting
Circa 2025
In Android Plaza, Harman explores the collapse of visual coherence through a process that begins with photographs of distorted television broadcasts—signals caught in the liminal space between presence and disappearance. These corrupted images are used to train machine learning systems, generating unstable compositions that flicker between figuration and abstraction. Into these already fractured outputs, Harman digitally paints, not to restore order, but to further destabilize the image—layering in new disruptions, artefacts, and visual misfires.
The result is a visual language that resists resolution. Figures emerge only to dissolve; faces blur into fields of interference; memory degrades into pure signal. The works do not depict in any traditional sense—they inhabit a space of aesthetic uncertainty, where perception itself becomes unreliable, and recognition is constantly deferred.
While the process engages with both AI and generative techniques, Android Plaza defies neat classification. It is not wholly generative, nor purely machine-made. It is a hybrid practice in which human intervention and machine error are locked in a recursive loop of distortion. Harman does not seek to correct the machine’s misremembering, but to amplify it—treating technological failure not as flaw but as method.
