Journal of
eISSN: 2573-2897
Submit manuscript...
Short Communication Volume 10 Issue 2
Why can’t artificial general intelligence be achieved?
Hong Sheng
Received: June 24, 2025 | Published: August 20, 2025
Citation: Sheng H. Why can’t artificial general intelligence be achieved? J His Arch & Anthropol Sci. 2025;10(1):43‒47. DOI: 10.15406/jhaas.2025.10.00323
Abstract
Agnosticism argues that human beings cannot fully comprehend everything in the universe, thus making it impossible to create a general artificial intelligence whose intelligence surpasses that of humans. The achievements of artificial intelligence so far have been limited to successes in specific, simple system domains. A truly “general” intelligence should be capable of simultaneously solving many problems (e.g., 100,000 tasks). Furthermore, when it comes to complex systems, their complexity far exceeds that of games, making artificial intelligence even less capable of handling them. The computational power of AI depends on computer speed, which has a fundamental limit the speed of light. Real-world problems often involve multi-dimensional scenarios whose complexity increases exponentially without an upper bound. Thus, finite computational capability cannot satisfy infinite computational demands. Human intelligence, distributed across various functions through natural evolution, is inherently more efficient than artificial intelligence at the individual level. Therefore, AI cannot fully surpass human intelligence, but can only outperform humans in certain specialized tasks. A conceivable future scenario involves a division of labor between humans and artificial intelligence, where cooperation and complementary strengths create a mutually beneficial world.
Keywords: artificial intelligence, agnosticism, complexity, computing speed, intelligent distribution
©2025 Sheng. This is an open access article distributed under the terms of the, which permits unrestricted use, distribution, and build upon your work non-commercially.