Armed Algorithms

摘要

Histories of artificial intelligence (AI) and robotics often overlook how physically crafted structures were crucial to the code that shaped their mutually constitutive evolution. This article explores these relationships by charting how the artificial intelligentsia at MIT and Stanford experimented with “armed algorithms”—robot arms interfaced to computers—in calibrated built environments, called microworlds, during the late 1960s and early 1970s. These manufactured spaces were designed to filter out the noise and recalcitrance of the real world that made robotic experimentation exceedingly difficult. Beyond their laboratory or scientific value, the microworlds also served as positive demonstrations of mutual orientation, which in turn recursively drove the mutual construction of the technoscientific Cold War world. To make their armed algorithms work, therefore, technologists on Route 128 and in Silicon Valley hacked not only their computers but also the physical and social world to push forward the frontiers of AI and robotics.