[Paper Review] To study the phenomenon of the Moravec's Paradox

Kush Agrawal|arXiv (Cornell University)|Dec 14, 2010

Action Observation and Synchronization5 citations

TL;DR

This paper investigates Moravec's Paradox, which posits that tasks requiring intuitive, sensory-motor intelligence—like walking or facial recognition—are far more difficult to implement in AI than abstract, logical tasks such as chess or complex calculations. The study demonstrates that while artificial intelligence excels at symbolic reasoning and computation, it struggles with perceptual and motor tasks due to the unconscious, evolutionarily ancient sensorimotor knowledge embedded in the human brain, which current AI lacks.

ABSTRACT

"Encoded in the large, highly evolved sensory and motor portions of the human brain is a billion years of experience about the nature of the world and how to survive in it. The deliberate process we call reasoning is, I believe, the thinnest veneer of human thought, effective only because it is supported by this much older and much powerful, though usually unconscious, sensor motor knowledge. We are all prodigious Olympians in perceptual and motor areas, so good that we make the difficult look easy. Abstract thought, though, is a new trick, perhaps less than 100 thousand years old. We have not yet mastered it. It is not all that intrinsically difficult; it just seems so when we do it."- Hans Moravec Moravec's paradox is involved with the fact that it is the seemingly easier day to day problems that are harder to implement in a machine, than the seemingly complicated logic based problems of today. The results prove that most artificially intelligent machines are as adept if not more than us at under-taking long calculations or even play chess, but their logic brings them nowhere when it comes to carrying out everyday tasks like walking, facial gesture recognition or speech recognition.

Motivation & Objective

• To analyze the fundamental discrepancy between human cognitive abilities and artificial intelligence performance across different task types.
• To investigate why seemingly simple, everyday perceptual and motor tasks are more challenging for machines than complex logical reasoning tasks.
• To highlight the role of unconscious, evolutionarily developed sensorimotor knowledge in human intelligence and its absence in current AI systems.
• To demonstrate that AI systems, despite advanced computational power, fail at tasks requiring embodied perception and motor control.

Proposed method

• The paper uses a conceptual and comparative analysis of human cognitive architecture and artificial intelligence systems.
• It draws on Hans Moravec's original formulation of the paradox to frame the analysis of AI capabilities and limitations.
• The study contrasts performance in symbolic reasoning tasks (e.g., chess, mathematical computation) with performance in perceptual and motor tasks (e.g., walking, facial recognition, speech recognition).
• It emphasizes the evolutionary development of sensorimotor systems in the human brain as a foundation for intuitive intelligence.
• The analysis relies on qualitative evidence from AI research and robotics, focusing on the gap between human-like performance and machine implementation.

Experimental results

Research questions

• RQ1 Why do artificial intelligence systems perform better at abstract logical tasks than at basic perceptual and motor tasks?
• RQ2 What explains the difficulty in replicating human-like sensorimotor skills in AI systems?
• RQ3 How does the evolutionary development of human sensory and motor systems contribute to the paradox?
• RQ4 To what extent does unconscious, embodied knowledge underlie human intelligence and hinder AI replication?

Key findings

• Artificial intelligence systems outperform humans in tasks requiring symbolic reasoning, such as playing chess or performing complex calculations.
• Despite their computational power, AI systems struggle significantly with basic perceptual and motor tasks like walking or facial gesture recognition.
• The human brain's sensorimotor systems, shaped by over a billion years of evolution, provide a deep, unconscious foundation for intelligence that current AI lacks.
• Abstract thought is a relatively recent development in human evolution—less than 100,000 years—making it less intuitive and more effortful to implement in machines.