Remixing AIs: mind swaps, hybrainity, and splicing musical models

Musical models created using machine learning techniques can be hacked, repurposed and spliced in many new ways. The current generation of models is not at any level of simulated consciousness sufficient to trigger immediate ethical blocks, and stands open to systems surgery. Beyond morphing of inputs and outputs, hidden layers of deep learning models may themselves be actively stimulated and substituted, from unit weights and biases to activation functions. Layers from multiple networks may be swapped and interpolated, from single units to complete layers. The hybridity of musical formation takes place at the level of model internals, in artistic transformations beyond standard transfer learning. This activity of AI code bending is dubbed here 'hybrainity', and alongside theoretical discussion of its origins, potential and ethics, examples of hacking particular machine learning models for new creative projects are provided, including applications in live performance and audiovisual generation.