The science-fiction nightmare usually arrives with thunder. Machines seize control, armies of robots march through ruined cities, and human beings wake to discover that their bodies have become components in a technological system they no longer control. Reality is rarely so theatrical. If something resembling The Matrix ever comes into existence, its origins may look considerably more mundane: a laboratory bench, a nutrient solution, some electrodes and a cluster of living human neurons quietly processing information.

That is what makes recent developments in biological computing so fascinating and disturbing. A project involving the National University of Singapore, data-centre company DayOne and Australian company Cortical Labs has assembled what is described as the world's first independently operated biologically integrated server rack. The prototype contains 20 CL1 biological computers, each incorporating about 800,000 lab-grown human neurons. That produces the headline-grabbing total of approximately 16 million neurons performing functions within a computing system.

There is an important qualification before anyone starts searching the basement for Hollywood's Keanu Reeves. These are not 16 million neurons cut from living people's brains, nor are 20 conscious human brains imprisoned inside a computer rack. The neurons are laboratory-grown cells derived through stem-cell techniques. They sit on microelectrode arrays that allow electrical signals to pass between the living cells and conventional digital systems.

Something conceptually extraordinary has nevertheless happened. We have crossed the boundary between using computers to simulate neurons and using actual living human neurons as components of computing systems.

At present the ethical distinction between this and The Matrix could hardly be greater. There is no good evidence that these small neural cultures possess anything resembling human consciousness, personal identity or the capacity for suffering characteristic of a developed human brain. To leap from today's biological computer to "scientists are enslaving conscious human minds" would turn a legitimate ethical question into science fiction presented as fact.

But that does not mean science fiction is irrelevant. Quite the opposite. The value of The Matrix is not that it predicted the CL1. It is that it poses a philosophical question that biological computing may eventually force us to answer: at what point does biological processing cease to be merely equipment and become something to which we owe moral obligations?

That question becomes considerably more interesting when we apply the slippery slope correctly. A slippery-slope argument is weak when it merely says that because A happened, Z inevitably follows. It becomes much stronger when we can identify incentives and intermediate steps connecting A to B, B to C and eventually to something that would once have seemed unthinkable.

The first step has already occurred. Living human neurons can be integrated with silicon and used experimentally for information processing.

The second is scaling. Twenty CL1 units contain roughly 16 million neurons, but there is no obvious commercial principle saying that biological computers must forever remain at that scale. If wetware proves useful, researchers will have powerful incentives to make biological neural networks larger, more stable, more interconnected and longer-lived.

The third is complexity. Researchers will want better learning, memory, adaptation and computational performance. Those objectives push biological systems towards increasingly sophisticated neural organisation.

The fourth is economics. Biological computing is being promoted partly because nervous systems perform extraordinary information processing on very little energy. The Singapore project's proponents explicitly identify energy efficiency as one attraction as electricity consumption by conventional data centres rises. Cortical Labs also envisages applications including drug discovery, humanoid robotics, cybersecurity and fraud detection.

Now imagine that biological processors eventually become economically useful. At that point the moral landscape changes because we are no longer discussing an eccentric laboratory experiment. We are discussing an industry. Industries seek scale, efficiency and standardisation. If a million neurons are useful, someone will ask whether ten million are better. If ten million perform well, somebody will experiment with a hundred million. If a particular biological architecture learns faster, laboratories will try to reproduce it. If particular genetic characteristics improve performance, biotechnology will provide incentives to optimise them.

The direction of travel becomes uncomfortable because the very characteristics we would want from advanced biological computers; learning, memory, adaptation, flexible problem-solving, overlap with characteristics that make nervous systems morally interesting.

There is no law of nature stating that greater neural complexity must produce consciousness. We do not understand consciousness well enough to make such a claim. Indeed, that ignorance is precisely the problem. We still cannot explain satisfactorily how the electrochemical activity of an ordinary human brain produces the subjective world of pain, pleasure, fear, memory, colour, thought and self-awareness.

Consequently, if future researchers construct increasingly complex human neural systems, how will they know when something morally significant has appeared? That is where the nightmare begins, not with an evil computer but with uncertainty.

Suppose a future biological processor contains hundreds of millions or billions of human neurons organised into sophisticated structures. Suppose it learns. Suppose it develops persistent internal states. Suppose it remembers previous stimulation. Suppose it changes its behaviour to avoid particular inputs. At what point does an electrical response become analogous to distress? At what point does avoidance become something resembling fear? At what point does information storage become memory in a morally relevant sense?

The obvious response is that today's systems are nowhere near this threshold. Quite so. But ethics is supposed to establish principles before technology reaches the dangerous threshold, not several years afterwards when enormous commercial investments depend upon declaring that no ethical problem exists.

There is an ominous precedent in humanity's tendency to define inconvenient beings according to their usefulness. Once something has economic value, there is a powerful incentive to classify it in whatever way permits continued exploitation. A sufficiently advanced biological computer would be extraordinarily inconvenient if somebody demonstrated that it could suffer. A billion-dollar wetware industry would have every reason to insist that its neural processors were merely tissue.

This leads to a genuinely Matrix-like possibility, although not quite the one portrayed by Hollywood. In the film, complete human beings are physically imprisoned by machines and incorporated into an industrial system. The real-world analogue might be subtler. Humanity might manufacture neural systems precisely for servitude. They would never have bodies, parents, childhoods or legal identities. They would exist because we wanted them to calculate, learn and work.

If such systems never developed consciousness, the ethical issue would remain limited. They would be biologically remarkable machines. But if some future version did become conscious, we would have created perhaps the purest slave imaginable. It could be born inside its prison.

It might have no muscles with which to struggle, no mouth with which to protest and no sensory world except the electrical signals supplied by its owners. Its entire environment could be controlled by software. Reward and punishment could consist of patterns of stimulation delivered directly into the neural network. Its continued biological existence would depend upon pumps supplying nutrients and technicians maintaining temperature, chemistry and sterility.

The remarkable thing is that pieces of this architecture already exist without possessing the sinister properties imagined here. The Singapore system connects living neurons to digital hardware through microelectrode arrays capable of stimulating and recording neural activity.

Again, this does not mean those neurons are conscious prisoners. It means that the engineering concept of connecting living neural tissue to a digitally controlled system is no longer science fiction. Push that technology forward fifty years and the philosophical questions become formidable.

The most disturbing scenario does not even require using an existing person's brain. Suppose biotechnology eventually allowed researchers to grow structures approaching the complexity of substantial portions of a human brain. They could conceivably be engineered for longevity, rapid learning or specialised cognitive functions. They might be connected into vast networks in which biological processors performed tasks for conventional artificial intelligence.

Then the boundary between AI and humanity would become extraordinarily difficult to draw. The silicon system might provide memory and communications while biological neural tissue supplied adaptive learning. Neither component alone would constitute the complete intelligence. The result would be neither ordinary computer nor ordinary human.

Who would own it? The corporation that grew the cells? The person who donated the original biological material? The programmers who trained the system? Could the biological component have rights of its own? Could it be switched off? Could unwanted memories be erased? Could millions of identical biological processors be manufactured from one cell line?

And there is an even darker possibility. If biological computing ever demonstrated advantages from using increasingly complete neural structures, pressure could eventually arise to obtain or preserve actual brains rather than merely grow simplified neural cultures.

There is no evidence that anybody involved in the present projects intends such a thing. The point of the slippery-slope argument is institutional rather than conspiratorial. Technology follows incentives. If biological intelligence becomes commercially valuable, there will be enormous incentives to discover how much biological intelligence can be engineered, preserved, connected and controlled.

At that point The Matrix ceases to look entirely ridiculous. The film got the engineering detail wrong. Human bodies make little sense as batteries. But perhaps its deeper idea was more interesting than its fictional energy economics: human biology itself becoming infrastructure. That is precisely the conceptual boundary biological computing has now begun to cross.

We should therefore resist two equally foolish reactions. The first is hysteria: declaring that 16 million tortured minds are already trapped inside a Singapore data centre. They aren't. The second is complacency: insisting that because today's neural cultures are primitive, no serious ethical question exists.

The sensible position lies between them. Biological computing may prove enormously beneficial. It could help researchers understand neurological disease, test drugs, investigate learning and perhaps develop radically more energy-efficient forms of computation. Those possibilities deserve serious exploration. The researchers themselves emphasise applications in neurological research and drug discovery as well as computing.

But exploration needs boundaries. Before biological computers become substantially more complex, researchers and governments should be asking what evidence would indicate morally relevant sentience, what levels of organisation deserve special oversight, how human-derived neural material should be classified, and what protections would apply if biological computational systems ever displayed credible signs of suffering or self-awareness.

We should also reject the dangerous assumption that something manufactured for a purpose can therefore never possess rights. Origin does not determine moral status. If humanity someday creates a conscious biological intelligence, saying "we grew it in a laboratory" would no more settle its moral status than saying "we own the server rack."

The Singapore prototype is therefore important for reasons extending well beyond whether biological processors eventually outperform silicon. Twenty machines containing 16 million living human neurons represent a small experiment by the standards of the human brain, but a considerable philosophical milestone.

For thousands of years humans built tools from stone, wood and metal. Then we built machines that manipulated information. Then we built machines inspired by the brain. Now we are beginning to build machines with living human neurons inside them.

Perhaps biological computing will remain safely on the non-conscious side of the moral boundary forever. Let us hope so. But technological civilisation has a poor record of refusing profitable possibilities merely because earlier generations would have found them disturbing.

The road to The Matrix would not begin with machines suddenly attacking humanity. It would begin much more innocently. Someone would discover that human neurons make useful hardware.

https://www.technocracy.news/world-first-biological-data-center-uses-16-million-living-human-neurons-to-process-data/