Why "AI consciousness" isn't coming anytime soon. (Anil Seth)
9 hours ago
- #Consciousness
- #Entropy
- #Artificial Intelligence
- An AI-operated jet bridge at Madrid-Barajas airport gets stuck in an infinite loop, requiring human intervention to resolve the issue.
- AI systems can get stuck in infinite loops due to fundamental limitations, unlike humans who adapt and solve problems dynamically.
- Alan Turing proved that no algorithm can always determine if another algorithm will stop or run forever, highlighting inherent limitations in AI.
- The 'frame problem' illustrates the difficulty AI has in distinguishing relevant from irrelevant information, leading to potential infinite loops.
- Human intelligence and consciousness are deeply tied to time and entropy, allowing us to avoid infinite loops and adapt flexibly.
- Living systems, unlike AI, are open systems that reduce entropy by exchanging energy with their environment, enabling survival and adaptability.
- Consciousness integrates survival-relevant information and guides behavior, making it crucial for avoiding repetitive, non-adaptive loops.
- Classical computation ignores the flow of time, making AI vulnerable to infinite loops, whereas biological systems are inherently time-bound.
- New approaches like neuromorphic computing and mortal computation aim to make AI more time-aware, but may still fall short of biological intelligence.
- Consciousness may be necessary for open-ended intelligence, as it integrates metabolic, neural, and experiential processes in a time-rich framework.