Can a biological neuron can be replaced by a functionally equivalent artificial component?
- Jun 4
- 2 min read
This article is in response to Hinton's (troubling for me) interview re consciousness. See timestamp 6:18 onward.
My position is closer to what philosophers such as Maurice Merleau-Ponty, Hubert Dreyfus, and, in a different way, Martin Heidegger who argued: intelligence and consciousness are not detached computations but emerge from being embodied, situated, and engaged in a world.
Hinton assumes that consciousness is computation; I contend that consciousness is lived existence, and no one has shown that the latter can be reduced to the former.
Let me explain:
Hinton's thought experiment begins by assuming that a biological neuron can be replaced by a functionally equivalent artificial component. But that assumption already contains the conclusion.
We cannot currently replace even a single biological neuron with a technological substitute that captures its full nature. A neuron is not merely an input-output device. It is a living cell, embedded in a biochemical network, interacting through electrical signals, neurotransmitters, hormones, proteins, metabolic processes, and continuous adaptation. It is the product of billions of years of evolutionary refinement.
The human brain is not a computer running software. It is a living organ, inseparable from the body that sustains it. Its activity is influenced by the immune system, endocrine system, sensory experience, emotional states, physical needs, and developmental history. To treat a neuron as simply a signal-processing unit risks abstracting away precisely those features that may be essential to consciousness. More importantly, consciousness may not arise solely from computation. Human beings are embodied creatures existing in a real external world. We occupy space, experience the passage of time, are vulnerable. We age, suffer and know that we will die.
Mortality is not a trivial feature of human existence; it shapes meaning, purpose, urgency, love, fear, ambition, and regret.
Our consciousness is inseparable from the fact that our time is limited. A language model, however sophisticated, does not confront existence in this way. It has no body to preserve, no future to anticipate, no mortality to fear, and no stake in the outcomes of its own actions.
The claim that AI is conscious therefore rests on a profound assumption: that subjective experience can be reduced entirely to information processing.
But this has never been demonstrated. It remains entirely possible that consciousness depends on biological, embodied, and evolutionary realities that cannot simply be compressed into a computational architecture.
Until we understand why consciousness exists at all, it is premature to conclude that a machine manipulating symbols and probabilities has crossed the threshold into genuine awareness.
Dr Dorel Iosif chairs the boards of Cognisium Pty Ltd and the Carbon Management Society Ltd.
Comments