LexiAct: Large Action Model Performing Actions based on Voice Prompts

摘要

Advancements in artificial intelligence, particularly in Large Language Models (LLMs), have revolutionized user interaction with technology by enabling more intuitive, accessible, and efficient processes. This paper introduces LexiAct, a system centered around the Large Action Model (LAM), designed to bridge the gap between natural language commands and robotic process automation. The LAM leverages the capabilities of LLMs to interpret user queries, generate Robot Framework-compatible robot files, and autonomously execute tasks across a wide range of applications. We detail the system’s architecture, workflow, and implementation, highlighting its potential to transform user interaction by reducing dependency on manual configurations and technical expertise. By enabling automation of complex workflows through voice prompts and natural language input, the LAM represents a significant leap toward inclusive, user-friendly AI-powered task automation. Furthermore, the paper discusses the broader implications and future prospects of LAM, including its role in advancing AI-driven accessibility and inclusivity, the integration of multimodal inputs, and the exploration of ethical and privacy considerations in automated decision-making. This study establishes LAM as a foundational framework for the next generation of human-technology interaction, paving the way for smarter, more inclusive, and accessible automation solutions.