Embodied AI: Desire, AI and Supernormal Stimuli

Embodied AI: Desire, AI & Supernormal Stimuli extends my earlier research on eroticism, censorship, and supernormal stimuli into generative AI. Rather than treating desire as an object to be represented, I approach it as an epistemological and behavioural force through which the assumptions embedded in computational systems can be exposed and tested.

The work establishes an experimental feedback loop between human physiological responsiveness and computational audiovisual generation. Instead of relying solely on linguistic prompts, the system uses physiological signals associated with the audience’s bodily responses as an input for modulating pre-existing films, images, sounds and sculptural elements, while simultaneously generating new audiovisual and spatial behaviours. The resulting changes in the environment feed back into the participants’ embodied experience, producing further physiological responses and continuously altering the subsequent computational output.

Operational System_Embodied AI: Desire, AI and Supernormal Stimuli

This creates a recursive process:

physiological response → computational modulation/generation → audiovisual stimulus → embodied response → physiological response → …

The artwork therefore treats the body not simply as a source of data, but as an active participant in an evolving human–machine system. The physiological signal functions as a form of embodied feedback through which the human participant and computational system continuously influence one another.

This research raises a question directly relevant to human–AI hybrid systems: what happens when physiological and affective responses become feedback signals through which an artificial system adapts its behaviour, while the human simultaneously adapts to the system’s responses? Rather than assuming that bodily signals transparently represent stable emotional states, the work deliberately foregrounds their ambiguity, circularity and susceptibility to contextual and audiovisual stimulation.

In this sense, the project provides an artistic and experimental foundation for investigating affective and social reinforcement in human–AI interaction: how an artificial agent might learn from embodied human responses, how humans modify their behaviour in response to an adaptive system, and how such reciprocal adaptation could produce emergent patterns of desire, attention and behaviour.

The project also extends my theoretical research on cognition and censorship. My doctoral research examined how affective and embodied processes can circumvent or counteract forms of conscious cognitive filtering. I now investigate how this relationship changes when affective responses become computationally legible and are incorporated into an adaptive human–AI feedback system. The central question shifts from how affect influences human cognition to how affect may become a mechanism through which humans and artificial agents mutually shape one another.