Human–Computer Interaction in the Agentic AI Era | by Chier Hu | Medium
Human–computer interaction (HCI) has historically focused on designing interfaces for deterministic computing systems. The rapid maturation of large language models (LLMs) and generative AI has created agentic AI: systems in which AI agents perceive their environment, plan, reason, and execute tasks autonomously. Unlike traditional software that waits for user commands or generative AI that merely produces content, agentic AI exhibits initiative and collaboration; it can decompose tasks, delegate to specialized subagents, monitor progress and adapt in real time. Recent prototypes (e.g., AutoGPT, AutoGen, OpenAI Agents) have shifted the HCI paradigm from single agents to teams of AI agents, creating multi‑agentic systems (MAS) accessible to ordinary users. This report synthesizes academic and industry literature up to April 2026 to analyze how HCI must evolve in the agentic AI era, highlight design challenges, and outline principles for building trustable, human‑centered agentic systems