Embodied AI in Machine Learning -- is it Really Embodied?

2025-05-15

AI Summary

This paper challenges whether current embodied AI truly achieves embodiment, arguing that mainstream large-model-driven robots exhibit only weak embodiment and inherit the disembodied limitations of Good Old-Fashioned AI (GOFAI). Method: It conducts the first systematic critique of “pseudo-embodiment,” situating embodied AI within an interdisciplinary framework integrating embodied cognition theory, behavior-based robotics, and modern multimodal learning to clarify core concepts and reexamine foundational paradigms. Through cross-paradigm analysis, it identifies three fundamental bottlenecks: (1) breakdown of the perception–action closed loop, (2) lack of endogenous world modeling, and (3) non-transferability of embodied experience. Contribution/Results: The paper proposes a pathway toward strong embodiment—centered on embodied agency, online embodied interaction, and evolutionary skill acquisition—providing critical theoretical criteria and strategic direction for foundational reconstruction and technical advancement in embodied AI.

Technology Category

• Intelligent Robots: Embodied AI
• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness
• Humans and AI: Other Foundations of Human Computation & AI

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Embodied Artificial Intelligence (Embodied AI) is gaining momentum in the machine learning communities with the goal of leveraging current progress in AI (deep learning, transformers, large language and visual-language models) to empower robots. In this chapter we put this work in the context of"Good Old-Fashioned Artificial Intelligence"(GOFAI) (Haugeland, 1989) and the behavior-based or embodied alternatives (R. A. Brooks 1991; Pfeifer and Scheier 2001). We claim that the AI-powered robots are only weakly embodied and inherit some of the problems of GOFAI. Moreover, we review and critically discuss the possibility of cross-embodiment learning (Padalkar et al. 2024). We identify fundamental roadblocks and propose directions on how to make progress.

Problem

• Examining if AI-powered robots are truly embodied
• Comparing Embodied AI with traditional GOFAI approaches
• Identifying challenges in cross-embodiment learning for robots

Innovation

• Leveraging AI progress for robot empowerment
• Critically analyzing weak embodiment in AI robots
• Exploring cross-embodiment learning possibilities