Agency-in-a-Box: Implications of Full-Stack Ai Agents for VCD Education Strategy

Abstract

Rapid development of artificial intelligence (AI) has shifted from passive, single-prompt-based generative models toward autonomous AI agent ecosystems (agentic AI). "Agency-in-a-Box" phenomenon illustrates the capability of full-stack AI agents, utilizing Large Language Models (LLMs) such Claude AI as workflow orchestrators via Application Programming Interfaces (APIs), analyze raw briefs, formulate visual strategies, and invoke external generative modules in an integrated manner. This potential disintermediation within the creative agency value chain poses serious challenges to Visual Communication Design (VCD) curricula that focus on technical instrument mastery. This qualitative research employs exploratory case study approach scenario study and curriculum gap analysis based, alongside structured interviews. Simulation result and interviews demonstrate that AI agents capable of accelerating brief synthesis and multimodal visual draft generation without significant conceptual degradation, critical human boundaries domain contextual empathy, local cultural semiotics, and ethical accountability. Result recommends reconstructing VCD curriculum based on: Meta-Design, Cultural Strategy, and Ethical AI Auditing.