Impossible Collaborations: Architecture and Machine Learning

1.5 units

The mass availability of Machine Learning processes makes individuals with limited experience of the tools perceive themselves as 'AI Experts'. The ever-evolving forms of the technology make it challenging to identify its vast potential influence on the design workflow. This course investigates how the presently available AI tools can productively disrupt prevailing design norms. Student projects will leverage Machine Learning's ability to disassociate from cultural affinities to defy preconceived notions of the architectural canon. These projects will imagine collaborations between canonical architectures and underrepresented, historically ignored architectures. We will rigorously speculate on how these hybrids become relics of a world beyond time, space, and identity by absorbing the exceptional qualities of their origins and introducing a novel method of interpreting precedents in architecture. We will explore AI processing tools like Google Colab, Automatic1111, RunwayML, Adobe Firefly, and Midjourney and delve into transposing/spatializing these images onto architectural elements using 3D modeling software like Rhinoceros, Blender, and Z-Brush. The final deliverables will be a series of estranged architectural models.