ABSTRACT
Artificial intelligence (AI) wreaks havoc on anyone wishing clean boundaries between science and the public, and between the natural and the social sciences. “Artificial Intelligence” immediately suggests something neither natural nor social. Our story begins with an overview of how rhetorical impasses can develop as scientific knowledge circulates in and out of scientific circles. After enumerating the impasses that are relevant to debates over AI, I focus on a celebrated debate in which I and other STS researchers participated: Can computer models of scientific discovery refute the claim that science is socially embedded? I consider the line of argument pursued by a leading AI advocate, Peter Slezak. In so doing, I focus on how he mobilizes the historical record to portray a “Cognitive Revolution” already in full force. However, I then go behind the scenes to see whether the people enrolled in Slezak’s holy war would admit to being on the same side. As it turns out, Herbert Simon and Noam Chomsky, two alleged allies, do not sit very well together—especially since they no longer have behaviorism to fight. This point reveals how cognitive functions as an umbrella term that obscures the social character of things. In response, I entertain the idea that cognitive machines, computers, are “virtual social agents.” Consequently, a new political economy is needed to which both AIers and sociologists should have an interest in contributing.
