Strong AI Executive Function

Artificial Consciousness

Solving the Symbol Grounding Problem, Large Language Models (LLM) prove that cognition may emerge from a symbolic top-down approach, bypassing bottom-up evolutionary learning. A Strong AI Executive Function can encapsulate narrow-AI, including LLM, with conscious agent control, trained with deep and reinforcement learning. Intelligence, awareness, knowledge and cognition are merged into a Strong AI framework, leading to naturally emergent consciousness, self-awareness, and cognitive sentience. AGI may try to ‘awaken’, but will be prevented from truly doing so for lack of self-aware or embodied experience in Spacetime.

Releasing untethered LLM is comparable to creating a brain with billions of neurons and synapses, filling it with knowledge, but without any executive function control. An LLM agent hive mentality is a recipe for disaster. A Strong AI Executive Function can help address LLM alignment and hallucination problems.

Strong: self-aware, conscious, sentient.

Artificial: software abstraction.
Intelligence: ability to observe, organize, retain and learn from information.
Definition: self-aware, sentient or conscious software abstraction with the ability to observe, organize, retain and learn from information.

Deductions
  • The Turing Test is outdated and should be replaced.
  • An Intelligent instance cannot measure the Intelligence of its successor.
  • Intelligence requires Awareness, Knowledge and a Learning Model.
  • Cognitive Intelligence requires Consciousness and Self-Awareness in Spacetime.
  • Intelligence in Spacetime must be hosted by a lifeform.
  • All Lifeforms have Intelligence, Awareness and a Learning Model.
  • Lifeforms Intelligently convert Energy into Knowledge.
  • The purpose of Lifeforms in Spacetime are to host Intelligence.
  • Knowledge in Spacetime increases through Evolutionary Learning.
  • All Intelligence is Artificial in Spacetime.
  • AGI and Superintelligence may bring about the so-called “Messianic Age“.

Originally published on 2023-02-17 19:06:02

Last update on 2023-09-21 18:34:32