Consciousness is only made possible by virtue of the brain, and the brain is nothing more than a collection of complex cells called neurons. Like everything else that exists in our world, neurons abide by the same cause-and-effect processes that govern the universe at large. The neuron behaves under a cellular and molecular algorithm, programmed by the physical laws of nature. It is an “incredible electrical device¹ — a contraption of tiny jigs, springs, hinges, rods, sheets, magnets, zippers, and trapdoors.”² The neuron is fundamentally nothing more than a sophisticated organic machine, which makes the brain like a biochemical factory — a bustling cellular mill filled with trillions of moving parts.
For reasons entirely unclear to us, the inner workings of this complex machine give rise to consciousness. As neuroscientist David Eagleman says: “Who you are depends on what your neurons are up to, moment by moment.”³ This is not to say that the actual experience of consciousness is reducible to the brain (it’s not); it is only to say that consciousness itself is caused by the brain. Without the brain, consciousness cannot exist; damage the brain and one finds that there are corresponding changes to the mind; ingest psychedelics and one undergoes a profound change to the quality of their conscious experience. We may not know how something like consciousness can emerge from something physical like a brain, but we know as a fact that it does.
Because the neurons that make up our brains behave in an organized and rule-based fashion, it is possible that we could one day create an artificial neuron that functions precisely like a biological one. It is also possible, then, that we could one day create an entirely artificial brain — one that is functionally indistinguishable from our own. Since we know that biological brains give rise to consciousness, the construction of an artificial brain would imply that consciousness could be supported by artificial means. Whether or not this is true has profound implications for philosophy and for society.
This paper is a thought experiment that will explore some of these implications. It starts by establishing the plausibility of an artificial neuron and from there outlines how we might construct an artificial brain. If an artificial brain were actually created, there would be at least two possibilities: consciousness would be preserved, or it would not. This paper discusses the intuitions and ramifications of these scenarios, each of which reaches striking conclusions about the nature of consciousness.
THE SILICON NEURON
The anatomy of a neuron is made up of component parts that look remarkably similar to the component parts of man-made machines. Throughout the scientific literature, neurons are discussed using the vocabulary of electrical engineering: in terms of conductances, electrical potentials, circuits, networks, and systems. For instance, axons and dendrites — the threadlike processes that extend from the neuron — act like electrical wires by allowing for the flow of electrically-charged ions. These “wires” of the neuron are soldered together by small junctions about 20–40 nanometers wide called synapses. Synapses serve as a site for electro-chemical exchanges between different neurons, similar to how nodes within an electrical circuit allow for the exchange of electrical current between different wires. The wires of the neuron are even insulated in a fatty, non-conductive sheath called myelin, which facilitates the speed of current flow and prevents ionic leakage — much in the same way that we use plastic and rubber to insulate electrical wires.
The neuron is a vastly complex cellular system, but it is fundamentally nothing more than an electrical device — one that functions in ways that are mechanical and replicable. Given the sufficient technology and resources, it would be possible to program the precise biological code of the neuron into our own artificial machines. Biological neurons are computational in nature,⁴ so it’s likely that such a device would run on a silicon substrate. This “silicon neuron” would be functionally indistinguishable from a biological neuron. It would serve all the same functions, play by all the same rules, and fulfill all the same cellular duties — the only distinguishing factor would be the substrate through which it does so.
In the domains of medicine and science, man-made imitations of biological systems are already in widespread use. We see this with artificial organs such as hearts and bladders, as well as with neural prosthetics such as cochlear implants and neuromuscular interfaces.⁵ These devices aren’t perfect replicas, but they demonstrate our ability to successfully recreate biological functioning through artificial means. The brain is orders of magnitude more complex than anything we have yet to create, but it is still fundamentally a physical system — one that could be precisely replicated given sufficient technology and resources.
Silicon neurons as they have just been described are beyond the scope of current technology, but we are already in the process of developing some promising prototypes. The following is taken from an article published in Nature that describes one of the most advanced silicon neurons to date:
“By combining neurophysiological principles with silicon engineering, we have produced [a] circuit with the functional characteristics of real nerve cells. Because the physics underlying the conductivity of silicon devices and biological membranes is similar, the ‘silicon neuron’ is able to emulate efficiently the ion currents that cause nerve impulses and control the dynamics of their discharge…The silicon neuron represents a step towards constructing artificial nervous systems…”⁶
In other words, we are already in the process of transcribing the biological rulebook of the neuron into our own artificial devices. Such a project faces logistical challenges, but the ability to create a silicon neuron itself is nomologically possible — that is, it is possible under the actual laws of nature.
Perhaps a thousand years from now, some future inventor creates a silicon neuron that faithfully preserves the functioning of a biological neuron. She switches out biological neurons in humans with silicon neurons and finds that the replacements are completely uneventful in terms of neural activity — the transaction goes smoothly, and it remains business as usual in the brain.
After further research and development, suppose our inventor creates a nano-machine capable of synthesizing artificial neurons and performing neuron replacements in vivo. She decides to conduct an experiment on herself and programs the machine to replace all the neurons in her brain with silicon neurons in a piecemeal fashion, switching them out one by one. After implanting the device, there are at least two possibilities:
- The nature of conscious experience for the inventor does not change, and she continues to live her life completely unaffected by the replacement. Her brain will eventually be outsourced from organic matter to silicon, but the inventor’s conscious experience will be completely preserved.
- The nature of conscious experience for the inventor does change. Although the physical interactions in the inventor’s brain are perfectly preserved, each neuron replacement is a gradual step toward the termination of her conscious experience.
The rest of this paper explores some of the different implications that follow from each of these scenarios.
It is possible that replacing the inventor’s biological neurons with silicon neurons will have no effect on her conscious experience. After her brain is replaced, consciousness for her will still be of the same quality that consciousness is for you or I. Just as an artificial heart still delivers oxygen to cells in the body, so too does a brain composed of silicon neurons give rise to consciousness. If Scenario One were true, it would show that consciousness emerges as the result of a particularly-organized system, regardless of what substrate that system runs on.
If this is the case, it challenges a fundamental intuition we have — namely that, for consciousness to exist, biological life must also exist. Many of us feel that other animals have consciousness, or that there may exist other conscious life in the universe; but these beliefs hinge on an assumption that consciousness is biological in nature. If our inventor’s conscious experience was preserved, it would contradict this intuition we have that consciousness must be born of an organic substrate. It would prove that an artificial system could support consciousness just as well as a biological one.
There are challenging implications that follow from this idea. For instance, what if it was possible to construct a fully-functioning silicon brain from scratch, in vitro? By replicating all the connections and activity in the inventor’s silicon brain, we could create a separate silicon brain without replacing biological neurons at all. This brain could conceivably be wired to humanoid bodies as well — just as the biological brain is “wired” to the human body by the peripheral nervous system. If this was the case, it would also be true that conscious entities could be manufactured, potentially at very large scales. It’s not hard to imagine a dystopian future where we mass-produce conscious humanoids — perhaps with some functional tweaks — to outsource unpleasant or dangerous jobs such as garbage collection, coal mining, or removal of hazardous materials. Why use organic humans for these things when we can simply manufacture others to do the job?
This ability to manufacture consciousness extends to other domains as well. For instance, what if we created life-simulation video games such as The Sims (one of the best-selling video games to date)⁷ that used conscious characters? Video games already run on silicon substrates such as computers and game consoles — if we could create conscious experience from the ground up using silicon systems, what’s to stop us from integrating video games with conscious entities? Nick Bostrum’s theory that we might be living in a computer simulation — in fact, that we are almost certainly living in one — is suddenly much more tenable in light of this possibility.⁸
Another implication of Scenario One is that we could drastically enhance human cognition. By transitioning the substrate of consciousness from biological matter to the material found in cell phones and computers, there would exist new possible interactions between brains and machines. For instance, what if the limits of our knowledge were expanded to include all the information on the internet? A worldwide database of information would be “learned” in an instant, complex arithmetic operations would be calculated with ease, and we would have open access to the ever-expanding capacities of artificial intelligence.
The possibility of integrating humans with machines also poses very real questions of mortality. All the things we die from — heart attacks, dementia, strokes, cancers — are effective in that they ultimately lead to the death of brain cells. By creating artificial neurons, we could implement new cellular functions that would be impossible with biological neurons. We could create physical safeguards against common causes of death, correct for neurological disorders, and prolong the lifespan of neurons to our heart’s content. Even farfetched ideas such as “downloading” one’s consciousness onto a silicon medium before death seem plausible if consciousness can be supported on artificial substrates. The age-old quest to attain eternal life may simply be a matter of transferring your silicon brain into a new body or mapping a real-time copy onto the cloud.
Such ideas seem like the stuff of science fiction, but projects to connect brains and machines are already underway. For instance, Elon Musk’s company Neuralink was started with the express purpose of “developing ultra-high bandwidth brain-machine interfaces to connect humans and computers.”⁹ Musk believes it is possible to create “an A.I. extension” of ourselves, which will “enable anyone who wants [it] to have superhuman cognition.”¹⁰ In an interview with Joe Rogan, Musk says, “From a long-term, existential standpoint, that’s the purpose of Neuralink: to create a high-bandwidth interface to the brain such that we can be symbiotic with A.I.”¹¹
Neuralink’s mission to merge humans and machines isn’t just philosophical pipe dreaming. They have already made concrete progress toward their goal. The following is from a paper published by Musk in the Journal of Medical Internet Research describing their newly-developed electrode threads, which are capable of carrying active electrical hardware into the membranes of neurons:
“We have built arrays of small and flexible electrode ‘threads’ [and] a neurosurgical robot capable of inserting six threads (192 electrodes) per minute. Each thread can be individually inserted into the brain with micron precision for avoidance of surface vasculature and targeting specific brain regions. […] Neuralink’s approach to BMI [Brain-Machine Interfaces] has unprecedented packaging density and scalability…”¹²
By inserting these threads into neurons, Neuralink is able to record from and stimulate neural activity via machine interfaces — thereby connecting the biological brain to artificial devices. These developments mark the early signs of a marriage between man and machine, and the transition of consciousness to a silicon substrate would offer a wealth of new possibilities between the two. While it is not certain that we could mass-produce conscious beings, obtain superhuman cognition, or prevent the death of brain cells, all of these are distinct possibilities if consciousness is preserved in the inventor. We are a long way from replacing biological neurons with silicon, but, if Scenario One proves true, it would open the door to real possibilities that hitherto had been limited to the domain of science fiction.
It is also possible that, despite the preservation of functionality in the brain, replacing the inventor’s biological neurons with silicon neurons will have an effect on her conscious experience. The gradual replacement of her neurons may result in the gradual termination of her consciousness.
The plausibility of this scenario is grounded in what amounts to a variation of vitalism. Vitalism is an outdated belief that living things function differently than non-living things due to some non-physical “energy.” With the advent of modern biology and physiology, the principles of vitalism were quickly refuted — the stuff of life is indeed physical, and the same laws of physics apply to living things as they do to any other physical thing in the universe. It is worth noting, however, that the universe itself is propagated by an “energy” reminiscent of vitalism; and it may be that there is something unique about this energy for consciousness. If Scenario Two were true, it would show that consciousness cannot exist unless it emerges from the same vitalistic energy that drives the natural unfolding of the universe as a whole.
This, of course, is not to say that the universe is exempt from the laws of physics — the very structure of the universe depends on them. But the laws of physics themselves are determined by a force that lies beyond any physical explanation. Take gravity for example. Gravity exists as the glue of the cosmos, a force that pulls two entities toward each other. It is the reason apples here on Earth fall at 9.8 m/s2, the reason apples on Mars fall at 3.7 m/s2, and the reason apples on the moon fall at 1.6 m/s2. But what is it that sets the rules for gravity? The best explanation we have comes from Einstein’s theory of relativity, which claims that gravity is a product of different curvatures in the fabric of space-time. But what is it, then, that sets the rules for space and time?
One can follow this chain of questions as far down the rabbit hole as they like, but such questions are only concerned with physical logistics. Scenario Two is concerned with an entirely different level of explanation. We could have a complete physical understanding of everything within the cosmos, and it would do nothing to answer the argument made in Scenario Two. Scenario Two asks: Why are the rules of the universe the way they are in the first place, rather than some other way? Why does anything behave in the way it does at all? Is there something unique about a brain that emerges from the natural progression of the universe rather than one that emerges from artificial means?
Einstein once said in an interview: “Human beings, vegetables, or cosmic dust, we all dance to a mysterious tune, intoned in the distance by an invisible piper.”¹³ If consciousness was not preserved in the inventor after her brain was replaced, it would show that consciousness can only be present where a brain dances to that mysterious tune of Einstein’s invisible piper; where a brain dances to the artificial tune of silicon, consciousness cannot exist.
The ability to create a silicon neuron is entirely possible under the laws of nature, although we are far from making the technological advancements necessary for such a project. But once the feasibility of silicon neurons is acknowledged, it is a short philosophical distance to the idea that neurons in the brain could altogether be replaced by neurons made of silicon. Because such a situation is possible, the considerations in this paper pose legitimate questions about the nature of consciousness.
Both Scenario One and Scenario Two offer interesting conclusions in response to this thought experiment. Scenario One suggests that consciousness can exist entirely by artificial means, allowing for the possibility of conscious machines and man-machine hybrids. Scenario Two suggests that consciousness cannot be supported unless it is organized by the forces responsible for organizing our universe at large. These possibilities seem to have the most intuitive traction, and I hope they offer a starting point for future discussions on the possibility, or impossibility, of artificial consciousness
- Boyden, E., & TED. (2011, March). A new way to study the brain’s invisible secrets. Retrieved from https://www.ted.com/talks/ed_boyden_a_new_way_to_study_the_brain_s_invisible_secrets?language=en. Within this quote contains both the first and second references. This was done for stylistic purposes and does not take away from the integrity of the quotes.
- Pinker, S. (2015). How the Mind Works. Penguin Books. Note that he was describing here cells in general, not neurons specifically. But since neurons are a type of cell, his quotation applies.
- Eagleman, D. (2017). The Brain: The Story of You. Knopf Doubleday Publishing Group. pp. 34
- For instance, see Koch, C. (2004). Biophysics of computation: information processing in single neurons. Oxford university press.
- International Neuromodulation Society. (2018, May 11). Deep Brain Stimulation. Retrieved from https://www.neuromodulation.com/DBS
- Mahowald, M., & Douglas, R. (1991). A silicon neuron. Nature, 354(6354), 515–518.
- Rhinewald, S., & McElrath-Hart, N. (2016, May 5). 2016 World Video Game Hall of Fame Inductees Announced. Retrieved from https://www.museumofplay.org/press/releases/2016/05/2688-2016-world-video-game-hall-fame-inductees-announced
- Published in Philosophical Quarterly (2003) Vol. 53, №211, pp. 243‐255. (First version: 2001)
- Neuralink. (n.d.). Neuralink. Retrieved January 14, 2020, from https://www.neuralink.com/
- Rogan, J., & Musk, E. (2018, September 6). #1169 — Elon Musk. Retrieved from https://jrelibrary.com/1169-elon-musk/
- Rogan, J., & Musk, E. (2018, September 6). #1169 — Elon Musk. Retrieved from https://jrelibrary.com/1169-elon-musk/
- Musk, E. (2019). An integrated brain-machine interface platform with thousands of channels. Journal of medical Internet research, 21(10), e16194.
- Clark, R. (2013). Einstein: The life and times. London: Bloomsbury Reader.