Big Questions and Small Minds: Conversational Interfaces, Evolution, and Language

Summary

This paper examines why conversational interfaces encourage people to perceive intelligence in computational systems, and how that perception affects the use of generative images. It relates the difficulty of distinguishing intelligence from its appearance to habits of perception and reasoning, arguing that conversation can obscure what these systems make possible as computational media. The paper proposes iterative approaches to image generation informed by visual practice and software development, with heuristics for constructing distinctive visual languages.