Infrastructural intelligence: Contemporary entanglements between neuroscience and AI

Johannes Bruder
DOI: https://doi.org/10.1016/bs.pbr.2017.06.004
2017-01-01
Progress in Brain Research
Abstract:In this chapter, I reflect on contemporary entanglements between artificial intelligence and the neurosciences by tracing the development of Google's recent DeepMind algorithms back to their roots in neuroscientific studies of episodic memory and imagination. Google promotes a new form of "infrastructural intelligence," which excels by constantly reassessing its cognitive architecture in exchange with a cloud of data that surrounds it, and exhibits putatively human capacities such as intuition. I argue that such (re)alignments of biological and artificial intelligence have been enabled by a paradigmatic infrastructuralization of the brain in contemporary neuroscience. This infrastructuralization is based in methodologies that epistemically liken the brain to complex systems of an entirely different scale (i.e., global logistics) and has given rise to diverse research efforts that target the neuronal infrastructures of higher cognitive functions such as empathy and creativity. What is at stake in this process is no less than the shape of brains to come and a revised understanding of the intelligent and creative social subject.
neurosciences
What problem does this paper attempt to address?