Playing Games with the Idea of Intelligence: AI, Game Studies, and the Ideal of Managerial Intelligence in Starcraft 2

• In Event: 4S at Play: Video Game Studies in STS - I

Abstract

This paper draws threads together from the social studies of artificial intelligence and machine learning research, game studies and the sociology of games and play to explore the link between games and the ongoing and currently hyperbolic project of solving the problem of general artificial intelligence (GAI) through deep learning and related techniques. Since the dawn of the GAI project in 1950s, games have been at the heart of both the technical imaginary and computational practices of AI researchers. Drawing on the work of Nathan Ensmenger, Lucy Suchman, Harry Collins and Martin Kusch in the social studies of AI on the one hand, and Simon Dor, Gary Alan Fine, and Julian Toglius in game studies on the other, I will examine the December 2018 defeat of TeamLiquid’s Grzegorz "MaNa" Komincz at the hands of Google Deepbrain’s AlphaStar in a human vs AI match of Starcraft 2. The fate of AI research and game culture have always been intertwined and I will review and build on some of the social and cultural dimensions of the entanglement of AI research with the game of Chess up to the 1990s, with the game of Go in 2015 and with contemporary videogame culture in 2018. At stake are the understanding of the combined technical imaginaries, software design practices and political consequences of the looming implementation of effective machine learning and artificial intelligence in many domains of social life.

Author

• Bart Simon, Concordia University, Montreal, Canada