LLMscape
Gottfried Haider, Jie Zhang
TL;DR
LLMscape investigates how humans and AI construct meaning under shared uncertainty by placing multiple large language model agents in a tangible, projection-mapped sandbox where participants reshape the environment. The project evolves through three iterations, advancing from a simple multi-turn simulation to generative agents with shared context and, finally, tool-calling and context engineering via the Model Context Protocol. By treating AI as embodied co-witnesses rather than deterministic tools, it reveals how meaning is co-constructed under epistemic constraints and prompts reflection on human-AI futures. The work contributes a methodological template for studying situated meaning-making in interactive AI installations and offers insight into how social, perceptual, and cognitive negotiation unfolds in shared spaces.
Abstract
LLMscape is an interactive installation that investigates how humans and AI construct meaning under shared conditions of uncertainty. Within a mutable, projection-mapped landscape, human participants reshape the world and engage with multiple AI agents, each developing incomplete and provisional accounts of their environment. Exhibited in Shanghai and continually evolving, the work positions AI not as deterministic tools but as embodied co-witnesses to an unstable world, examining the parallels between human and artificial meaning-making and inviting reflection on our shared epistemic limits.