Games, AI, and Systems

Michael Straeubig's 2019 essay in Eludamos examines the relationship between games and artificial intelligence, arguing that the two fields have developed largely in parallel and that game-based perspectives offer a more grounded way to think about AI than prevailing speculative debates. The paper contrasts the 2002 rhetoric of an "AI revolution" with the present landscape, contending that recent progress rests on decades-old algorithms made practical by dramatic declines in computing cost. The author cites the falling price of computation, with the cost of a teraflop dropping from an inflation-adjusted US$156.8 trillion in 1961 to roughly US$30 in 2017, alongside the availability of large datasets. He also acknowledges mounting criticism of AI concerning replicability, bias, opacity, and fragility.

The central contribution is a framework of seven roles that AI can play within games. These include the Mechanic, whose invisible logic structures gameplay; the Alter/Ego, which serves as the player's counterpart; the Observer, which performs player analytics; the Protector, which enforces anti-cheat rules and addresses harassment; the Player, exemplified by agents trained through deep reinforcement learning on titles such as Atari games, Go, and StarCraft II; and the Creator, which generates content procedurally, as in No Man's Sky. The framework culminates in a speculative seventh role, the "God," an AI integrating all the others. Straeubig argues this is technically plausible without human-level consciousness and may emerge from persistent virtual worlds rather than isolated laboratories.

The paper's scope is primarily conceptual and historical, drawing on AI research, game studies, and industry economics. Its speculative conclusions about a God-like AI are offered as argument rather than empirical finding, and the author calls for transdisciplinary research to shape this trajectory. Much of the closing material consists of references situating the work within scholarship on technological singularity, the historical divide between game AI and academic AI, and systems theory.