General Artificial Intelligence needs Consciousness

Summary

Creating human-like artificial general intelligence that is genuinely creative and sets its own goals may require consciousness, which machines cannot achieve through conventional computing. Human creativity stems from conscious experience and stochastic processes that computers can only simulate, not replicate. The frame and halting problems also limit traditional systems, while consciousness can overcome such boundaries. Therefore, sentient AGI is not possible with current computer science, though alternative approaches like biocomputing or quantum models might eventually lead to sentient machines, raising ethical concerns.

Study at a glance

Abstract

Human-like, Artificial General Intelligence (AGI) must autonomously explore reality, set its own goals, and demonstrate genuine creativity, such as making original scientific discoveries. Creativity in humans is rooted in high-level consciousness and the subjective feeling of personhood, making it unattainable for machines that are merely intelligent but not conscious, because creativity arises from stochastic processes machines can only emulate, not reproduce it. The frame and halting problems further limit conventional computational systems, preventing them from transcending their operational boundaries. Consciousness, on the other hand, can overcome operational boundaries set by evolution and access deeper layers of reality. Consequently, sentient AGI is not possible using conventional computer science approaches. Alternative approaches—such as biocomputing or quantum-based models of consciousness—might yield sentient machines in the future, though such hypothetical advancements raise profound ethical questions about creating new forms of awareness.

Tags