About                   Projects                   Research                Consultancy


Polymorph developed through distinct but complementary questions brought by each member of the group. These approaches formed a shared research environment around AI, anchored in prior research trajectories.

Johnny Golding’s philosophical framework of emergence, developed through her broader work on post-Newtonian analytics, provided a philosophical orientation. Her account of non-instrumental poietic praxis as a way of investigating distributed synthetic systems, together with her investigation into the possibilities opened by systemic incompleteness, shaped the lab’s line of inquiry into the systemic unknowns.

Maggie Roberts’ long-term research with 0rphan Drift on the distributed transformations of cephalopod skin emphasised the logics of pattern, shifting figure-ground relations, and the sensorial instability of underwater environments in which octopus bodies move continuously across multiple states. This research informed the design of complex adaptive dynamics in AI systems attuned to physical conditions.

Jeremy Keenan’s research on dither, the productive role of noise, and the translation between sensory modalities through physical computing contributed to the project’s approach to sound and real-time generative environments. His expertise in TouchDesigner and Max/MSP shaped the treatment of sound, image, sensor input, and feedback as active elements within the training cycle and the wider activity of the AI system.

Sonia Bernac’s research into data mixtures, synthetic misalignment, and complexity science shaped the lab’s approach to system-level change. Her work on heterogeneous AI infrastructures informed the design of systems in which image, sound, sensor data, and physical processes could enter shifting relations across the training cycle. Her work on phase change further contributed to the project’s understanding of self-reorganisation and shifts in systemic mode, or mood.

Across these different lines of inquiry, the research converged around camouflage, the productive use of noise, incommensurable systemic logics, and the conditions under which systems change.










Maggie Roberts’ contribution drew on her practice with 0rphan Drift, the collective she co-founded in London in 1994. Across video, animation, text, installation, and digital processes, 0rphan Drift has explored unstable perception, machine vision, speculative fiction, and forms of embodiment that exceed the single human observer.

Roberts brought this trajectory into Polymorph through her research into interspecies seeing, cephalopod perception, camouflage, figure-ground instability, and distributed surface transformation. Her work with octopuses attends to bodies that shift through colour, pattern, texture, movement, and environment, producing forms of visibility that are temporary, adaptive, and difficult to separate from their surroundings.

Within Polymorph, this informed experiments with reflective surfaces, sensors, image feedback, sound, and environmental disturbance. Her work helped articulate the system as a responsive material environment in which patterns emerge through changing relations between computational processes, physical matter, and surrounding conditions.


Johnny Golding led the project  as Principal Investigator, bringing to the project her philosophical work on emergence, post-Newtonian analytics, non-reductive logic, and attunement. Her account of emergence provided an important intellectual orientation for the lab’s shared inquiry into synthetic environments, systemic uncertainty, and unfamiliar forms of material intelligence.

Golding’s work on non-instrumental poietic praxis informed Polymorph’s methodological approach. Making was treated within the project as a rigorous mode of thought, and AI as a distributed synthetic environment whose behaviour could be entered, tested, and reconfigured through practice. This helped position the project as an investigation into how artificial systems generate forms of relation, response, and material intelligence.

Golding’s engagement with incompleteness, including Gödelian logics of non-closure, shaped the project’s resistance to totalising models of artificial systems. Polymorph was developed as a system that would not fully settle into a fixed state: matter, code, signal, perception, and environment remained in active relation, allowing emergent patterns, recursive disturbances, and new modes of organisation to appear.











Sonia Bernac’s contribution drew on her research into heterogeneous synthetic infrastructures, incommensurability, and complexity science. She approached AI systems as expanded synthetic corpora: distributed  bodies in which data mixtures, model organisation, code, signal, substrate, and generated output are coupled without becoming commensurate. Within Polymorph, this meant treating misalignment between image, sound, sensor input, and physical process as a generative condition of behaviour, with partial translation enabling incompatible elements to form a synthetic ecology without resolving into a shared logic.

Her work on dataset re-folding informed the project’s approach to relations between datasets as shifting under changing conditions, rather than fixed by inherited  training partitions.

Her engagement with complexity science and multi-agent dynamics brought the question of synthetic self-organisation into Polymorph: how an artificial environment might self-assemble or alter its internal dynamics. Phase change offered a way of understanding such regime shifts within the same system, with multistability marking the system’s capacity to sustain more than one stable organisation, and changes in texture, rhythm, and aesthetic form making those transitions perceptible.






Jeremy Keenan’s  contribution to Polymorph drew on his practice with sound, motion, feedback, data, sensors, and light. His work moves across generative audiovisual composition, sonic spaces, latent spaces, kinetic artworks, compound physical/virtual worlds, and hybrid media environments, with particular attention to real and imaginary signals, audible and inaudible feedback, and the lateral misuse of generative AI.

Keenan brought this trajectory into Polymorph through his research into dither, noise, physical computing, and the translation between sensory modalities. His experience with TouchDesigner and Max/MSP was central to the project’s construction. The whole system was orchestrated in TouchDesigner, allowing image, sound, sensor input, feedback processes, and environmental data to be routed, combined, and reconfigured in real time.

Within Polymorph, this informed experiments with sound, vibration, signal processing, live feedback, and sensor-based response. His work helped articulate sound as an active computational and material process within the training cycle and the wider activity of the AI system.