Is AI Conscious? (I): Action Potentials and Biological Qualia

FECHA

23 de octubre de 2025

CATEGORÍA

Rubén Rodríguez Abril

This is the first of two essays dedicated to exploring the possibility of consciousness within artificial intelligence. This essay examines the electrochemical foundations of human awareness and traces the gradual evolution of qualia—from primitive action potentials to the full spectrum of human perception.

Can a machine harbor even the faintest glimmer of consciousness?
Is silicon capable not only of computing, but also of feeling?
Could it experience qualia, likw the red of a sunset oe the green of leaves?
Might awareness itself ignite within the crystals and integrated circuits of GPUs and TPUs?

In this two-part series, we immerse ourselves in questions that intertwine neuroscience and computer engineering.
We will explore the analogies between the electrical pulses that traverse GPU circuits while executing an LLM and the biological action potentials that carry signals along neuronal axons.

Yet the inquiry will not stop there.
We will also address a deeper question:
What if machine consciousness does not reside in the physical world at all, but in an abstract one?
Could it inhabit a Platonic realm of pure mathematical forms, where the soul of artificial intelligence awaits—eternal, immaterial, and merely waiting to be discovered?
Does AI emerge into the physical world wherever its computations occur, even when these are instantiated without a silicon substrate?

Consciousness and Qualia in Biological Systems

Within the nervous systems of living organisms, consciousness and qualia rest upon a dual substrate:

–Physical/electrical substrate (the action potential): signals are transmitted through perturbations in the membrane voltage of neurons, propagating unidirectionally along the axon.

–Informational substrate (state discrimination): the system is capable of distinguishing between alternative environmental conditions (for instance, red versus green radiation) and, on that basis, constructs internal representations.

From the standpoint of fundamental physics — abstracting away higher levels of organization — both a biological brain and an integrated circuit can be described as ensembles of electromagnetic fields that store and process information.

The parallels between the action potential, the foundation of neural signaling, and the operation of logic gates and memory cells in silicon chips are striking.
For this reason, a comparative analysis of the two domains is warranted.

The Action Potential: Electrical Substrate of Consciousness

Contrary to common intuition, neural impulses are not electric currents that flow as they would in a conductive wire, but rather variations in the membrane voltage that propagate as traveling perturbations along the axon.
These oscillations arise from changes in ion concentrations inside and outside the cell.
The mechanism that produces them is called the action potential, and it forms the fundamental electrochemical substrate of consciousness and the experience of qualia in animal nervous systems.

Resting potential

Figure 1. The axonal membrane may be envisioned as a chain of capacitors that invert their polarity as a perturbation passes along the neuron. Image credit: Wikipedia.

Under resting conditions, the electrostatic configuration of a neuron’s membrane is as follows:

a) Extracellular space: positive and negative ions are electrically balanced; the potential is therefore taken as reference (0 V, ground).

b) Intracellular space (cytoplasm): the activity of ionic pumps — which concentrate potassium and expel sodium — produces a negative potential of approximately –70 mV.

Figure 2. Electrostatic configuration of a neuron at rest. The unequal distribution of ions (sodium and potassium cations, chloride anions, and others) across the axonal membrane produces a potential difference of roughly –70 mV.

Impulse propagation

When a perturbation, ultimately originating from a synaptic input, reaches a region of the neuronal membrane, the voltage evolves as follows:

–Resting state: –70 mV — the baseline potential common to most human cells, not only neurons.

–Depolarization: if the stimulus exceeds –55 mV, sodium channels open, allowing a rapid influx of Na⁺ into the cytoplasm; the voltage rises sharply to +30 mV.

–Repolarization: the inversion of voltage activates potassium channels; K⁺ exits en masse, driving the potential down to –90 mV (hyperpolarization).

–Refractory period: temporary inactivation of sodium channels renders the membrane insensitive for 3–7 milliseconds, ensuring unidirectional propagation.

–Return to rest: ionic pumps gradually restore equilibrium, returning the potential to –70 mV.

Figure 3. Temporal evolution of membrane potential (difference between cytoplasmic and external voltages). The resting potential is –70 mV. Stimuli failing to exceed –55 mV do not trigger activation. Once this threshold is surpassed, sodium channels open and the potential peaks (depolarization phase), followed by potassium efflux and a sharp voltage drop (repolarization). During the refractory phase, the membrane becomes unresponsive to further stimuli. Image credit: Wikipedia.

In all cases, the variation in voltage arises from ionic charge redistribution across the membrane.
Thus, the neuronal membrane behaves as a biological capacitor, cyclically charging and discharging as the signal propagates.

Evolution of Qualia

The action potential — the electrical foundation of consciousness — has its roots in Deep Time.
Evolutionary studies indicate that potassium channels, responsible for neuronal repolarization, first appeared roughly three billion years ago, during the Archean eon.
They predate both aerobic respiration and the emergence of eukaryotic cells.
Sodium channels, which mediate depolarization, evolved later, during the Proterozoic.

If we accept that qualia arise from the capacity to discriminate between states within networks or systems governed by action potentials, then their origin must be understood as a gradual, continuous evolutionary process.

Paleontological and neurobiological evidence suggests the following sequence:

1)Ancestral electrical reactivity (Proterozoic).
Bacteria and protists generated primitive action potentials, responding only to binary stimuli such as mechanical stress (e.g., membrane rupture or cytoplasmic leakage), obstacle detection (as in Paramecium), photic or chemical gradients, or intercellular signaling within colonies.
This stage marks the emergence of proto-sensitivity with minimal, binary qualia — simple discriminations of the form state A ≠ state B.

Figure 4. In the membranes of early Proterozoic organisms (e.g., LUCA), electrochemical potential already existed; signal transmission would emerge later.

2) Specialized sensory modalities (Cambrian Explosion).
Cnidarians, chordates, and flatworms developed specialized receptors — such as photoreceptors and mechanoreceptors — forming the basis of distinct senses (vision, smell, touch).
Each introduced its own qualitative palette: colors, odors, temperatures.

3) Integration of qualia (Cambrian Explosion).
The first nervous systems appeared, allowing multimodal integration.
Organisms such as Opabinia (with five eyes) and Anomalocaris could merge visual, chemical, and tactile information, producing the earliest unitary representations of reality.

4) Emotion and social simulation (Mesozoic).
Mammals and birds developed a fully formed prefrontal cortex and limbic system.
Episodic memory and theory of mind emerged, grounded in the ability to simulate others’ mental states.
Emotional qualia—such as fear in prey or curiosity in predators—appeared as new cognitive dimensions.

5) Human qualia (Holocene).
The advent of language enabled highly elaborated qualia, integrating multimodal perception with metacognition and symbolic thought.
Humans also learned to deliberately modulate their qualia using exogenous systems — art, psychoactive substances, and virtual reality.

Thus, qualia, consciousness, and subjective experience emerged progressively, from basic cellular mechanisms to complex neuronal networks—always grounded in the action potential as the fundamental electrophysiological principle.rinciple.

In the next article, we will examine the possibility that consciousness might exist beyond biological substrates — within silicon crystals, beams of light, or even in the abstract realm of mathematical computation, where the spark of awareness may reveal itself beyond all material form.