“Then God said, “Let us make humankind in our image, according to our likeness; and let them have dominion over the fish of the sea, and over the birds of the air, and over the cattle, and over all the wild animals of the earth, and over every creeping thing that creeps upon the earth.” (NRSV 1:26)
There is an argument that the latest technology in the world has seen humanity play God for some time now. We have shaped the natural landscape, bending it to our will, beyond all recognition. Of course, we are not the only living creature whose presence has an impact on the world around it and geography has always been hugely influenced by the biology it supports. But the extent of our footprint is without parallel.
The trouble is, while we take on many of the creationist and management responsibilities of a God, we lack the all-seeing omnipresence of the God of Abrahamic religions. So when humanity manipulates nature to its immediate advantage, we tend to do so without a great deal of appreciation for the long term and ripple-effect consequences.
Now, in 2019, humanity stands at the brink of perhaps its greatest achievement – the tantalising prospect of unlocking the secrets that might allow us to synthetically create biological organisms ourselves. And not just simple, one cell or primitive organisms but complex genomes. One day in the foreseeable future we might even be able to synthetically reproduce, and learn to edit, our own human genome.
And separately to these developments in biotechnology, breakthroughs in the latest technology in the world of neural network-based AI raises the question of if and when we might be able to accurately or practically replicate the human mind through code. Or a combination of code and biotechnology.
How likely, and close, is the possibility of us taking the next step towards become the God that made us in his image? Will we one day have the knowledge and tools to create our own beings in our image?
5 years ago in 2014 Romesberg Lab, the Californian research centre of renowned biochemist Floyd Romesberg, created something called xeno nucleic adic, or XNA. XNA is essentially a synthetic alternative to DNA, built from amino acids that do not naturally occur.
In 2016, Synthetic Genomics, another California-based research centre specialising in synthetic biology created an artificial lifeform named ‘JCVI-syn3.0’. It had a 473-gene genome. The private company’s team of scientists, led by Craig Venter, created JCVI-syn3.0 in an effort to gain deeper understanding of the function of individual genes, from the creation of RNA and proteins, the preservation of genetic consistency, or fidelity, across reproduction and the creation of cell membrane. It’s still not clear what around a third of the genes in JCVI-syn3.0 do.
And over the past couple of years there have been leaps forward in our skill at using the biotech gene-editing tool CRISPR. It allows us to snip the genome where we choose and substitute or add DNA sequences to it. CRISPR technology has already been controversially used in China by scientist He Jiankui. The first genetically modified human babies were born last year after he modified the genome of their embryos to ensure the HIV virus carried by their father could not be passed on to them.
The combination of these three developments raises the real possibility that we will soon have the ability, whether or not it is acted upon or how, to create artificial life. Or artificially edit natural life, turning it into something new.
The positive of these biotechnology developments is that we will almost certainly soon be able to correct the kind of genetic defects that blight the lives of those who suffer from them. The bigger picture, and the one that holds many unknown potential consequences and ethical and philosophical questions, is that we will also have the power to create new life forms that have never existed in nature.
We are still a long way away from being able to synthetically create multicellular life forms able to sexually reproduce themselves. But we have undoubtedly now taken the first steps along that road. And history tells us that once the Pandora’s Box has been opened, our thirst for knowledge and understanding will mean it is almost impossible for the lid to be closed again. Research will inevitably continue and the final destination, unless experiments are put a stop to which may be unrealistic in practical terms now the tease to the greatest scientific intellects is out there, would be artificially evolved humans.
The Human Mind Recreated By Artificial Intelligence
Some scientists believe we will never possess the capacity to truly unravel the vast intricacies of how the human mind works in all of its minutiae. The concept is explained by a quote from a neuroscientist in a Financial Times article exploring the possibility of AI ever reaching the point it can be considered a code-based equivalent to the human mind.
“You cannot cut butter with a knife made of butter”.
The theory rests on the assumption that the workings of the human brain can be replicated by a computer is just the latest in a long line of false metaphors we have relied upon throughout history in our attempts to wrap our brains around our brains. In previous eras the workings of our brains were compared to hydraulic mechanisms and then steam engines and telephone exchanges. Presuming its workings can be replicated by computer hardware and code is nothing more than a reflection of the computer’s dominant role in the productivity of contemporary society. But at the end of the day, is this not little more than another metaphor we are drawn to in the vain attempt to understand mysteries that will intrinsically remain elusive to us?
The human brain consists of around 85 billion nerve cells interconnected by 150 trillion connections. It is said that the organ is the most complex structure in the known universe. However, those who favour the computer metaphor say that despite the mind boggling number of nerve cells that make up our brain, they are all simple input/output devices that process data, whose input is in the form of electrical stimulation and output again electrical impulses – just like the electrical switches in a computer.
However, while we might think that this interpretation of how the human brain works would mean it is only a matter of time before we can create a computer just as complex, a huge chasm remains to be bridged. How does this storm of electro-chemical activity spread across space and time generate ‘conscious’ and ‘coherent’ experience?
If we accept that the brain, and human mind, is a purely and ephemeral physical system and discount the mysterious intangible non-physical concept of a ‘soul’ that somehow interacts with the physical, the way the brain works must be subject to the laws of physics. It logically follows that the human mind would then, in theory at least, be ‘computable’. This raises the practical question of how close we are to understanding how the human brain works and if AI can help us understand it better as well as deeper understanding leading to the creation of better AI?
Is It A Fundamental Mistake to Separate Mind and Body?
Another complexity to our attempts to understand the workings of our minds lies in their relationship to our bodies. There is a strong argument that the ‘thinking’ achieved by our brains is intrinsically tied to their ‘interface’ – the body they are attached to. The raw data that brains process comes from our bodies and in turn reach our bodies from their external environment. Without this data the brain has nothing to ‘learn’ from, leading some to the conclusion that thought distinct from embodiment is not possible.
One theory, among the most plausible, is that our brains are prediction processors. They help us navigate the world in search of food, shelter and sex. A map of our environment is created in our brains. This predictive theory goes a long way to explaining our sense of confusion if we encounter one more or less step than we expect and the existence of optical illusions.
Are Our Brains Just Too Complex For Us To Ever Be Able to Understand Them Fully?
Perhaps the bottleneck to an AI that truly matches the functions and processes of the human mind is the incredible complexity of our brains. The latest technology in the world of MRI scans allows us to look at a brain down to the scale of a single cubic millimetre. However, that cubic millimetre of brain can contain a 100,000 neurons and a billion synapses, arranged in a highly specific combination in relation to billions and trillions of other highly specific combinations. Can we truly expect to ever recreate a computer-based model of such intricacy or even first gain insight into its granular details?
To date one complete brain connectome has been established – that of the simplest tapeworm. It consists of just 307 neurons and 7,000 synapses and took years. A connectome is also just the most superficial layer of an entire nervous system. It can be compared to a map of a city and hoping to extrapolate an understanding of that city in all of its infrastructure, sights, sounds, smells and activity. We similarly lack any clear idea of how the map of the simplest of brains relates to senses and instinctive behavioural patterns before even considering how much more complex the relationship is to the thoughts and feelings a human brain creates.
Could AI Help Us Understand The Human Brain?
While the most cutting edge AI algorithms can do extraordinary things, they are limited to being able to perform one task they are programmed to. The ‘neural networks’ that AI is modelled on work like a brain in a very ‘big picture’ way. They are layered, with each layer handling distinct, related data inputs. For example, in the case of a driverless car algorithm, one layer will process the edges of road signs to determine shape, another colour a third design.
The output of these layers is combined in a way that means the algorithm can identify a particular road sign, even if it is partially obscured by a branch or fog, which would be processed and added into the mix by additional layers still. The output of the different neural network layers and in combination, feeds back into the algorithm and modifies future input in a way described as ‘learning’. But it is not really how our brains learn. AI algorithms only process input and output in a loop, leading to patterns of statistical association and classification. Not explanation or understanding.
General Intelligence AI that can perform numerous, unrelated tasks is still a pipedream. And it looks like the current neural networks approach to AI will never achieve this. One answer could be research around the concept of ‘neuromorphic chips’ as an alternative to neural networks. Neuromorphic chips are designed to resemble brain nerve cells physically, rather than as a code-based representation. However, recreating the human brain from such chips would necessitate understanding its deep structure far better than we currently do.
Is it, however, possible that future developments in neural network and perhaps new forms of AI, with their incredible processing power and ability to isolate patterns far beyond the scope of what the human mind’s speed and ability to retain and consciously combine data is capable of, could one day unravel the specifics of the billions and trillions of elements that combine in our brains?
Perhaps, but there is no prospect that day is anywhere on the horizon. Fears and excitement around the concept of ‘singularity’ – machines that are conscious entities in a way practicality indistinguishable to ourselves – might make for an interesting read. But the practical reality is that this will almost certainly not happen in the lifetime of anyone alive today, probably during the lifetime of several subsequent generations and quite possible never.
Will Synthetic Lifeforms Exist In Our Lifetimes?
If synthetic lifeforms do come into existence over the next decade, they will almost certainly be biologically based and an evolution of the kind of research that led to JCVI-syn3.0. It is far simpler to stimulate the beginnings of life, or edit them, with an organism subsequently developing through processes we don’t completely understand than it is to synthetically create the finished article.
The repercussions and questions around new synthetic or artificial forms of life, or synthetically evolved life forms that result in us editing existing genomes, are the more pressing ethical and philosophical debate. That is a reality that is on the horizon. Building a mind that compares to our own isn’t.