Can artificial intelligence reach human thought?

Area of Science:

• Cognitive Science
• Artificial Intelligence
• Neuroscience

Background:

• Deep learning advancements prompt debate on AI surpassing human intelligence.
• Existing definitions of AI intelligence focus on task completion, not human cognition's essence.

Purpose of the Study:

• To critically evaluate the claim that AI can reach and surpass human thought.
• To explore the limitations of current AI in replicating human consciousness and understanding.

Main Methods:

• Analysis of AI definitions of intelligence.
• Comparison of machine and brain understanding.
• Consideration of embodied cognition, brain lateralization, and glial cell function.
• Application of Gödel's incompleteness and Turing's results to consciousness.

Main Results:

• AI's goal-oriented definition of intelligence is insufficient for human thought.
• Human consciousness involves subjective experience, embodiment, and unconscious processes, which AI currently ignores.
• Gödel's and Turing's theorems suggest consciousness exceeds computational limits.

Conclusions:

• AI, by mimicking only parts of the cerebral cortex, fails to capture the richness of human consciousness.
• The transition from unconscious to conscious experience involves information loss, a complexity beyond current AI.
• Embodied cognition, brain lateralization, and glial cells are crucial for human thought, factors often overlooked in AI development.