The Signal in the Noise: AI, Games, and the Art of Denoising

Dylan Goldblatt, Ph.D., AI Strategist and Applied Researcher, Kennesaw State University

About the Talk

From AI-generated voices to procedurally crafted game worlds, noise becomes the fundamental raw material from which meaning emerges. This talk examines the ways modern AI systems, including large language models and diffusion-based graphics engines, harness noise and refine it into structured, playable, and engaging experiences. Through connections to text-to-speech synthesis, music generation, procedural game design, interactive narratives, and player behavior analysis, the discussion presents denoising as a unifying principle across these fields. Noise enables variation, creativity, and complexity, while AI systems extract signal by clarifying voices within randomness, generating coherent worlds from stochastic inputs, and identifying meaningful patterns within chaotic player interactions. Viewing AI through the interpretive lens of noise and denoising reveals how contemporary generative systems shape the boundaries between order and chaos, structure and serendipity, in the future of games.