By Keith Hayward
Keith is a committee member of Farnham Humanists. He is a retired motor engineer and currently runs a retail garden nursery in Bisley, Surrey. In this article, he reflects on whether, in the centuries to come, artificial intelligence robots will replace us or be our partners.
Artificial intelligence is a relatively new science. The foundations for it were laid down in the
1940s by pioneers including Alan Turing. At that time, transistors had not been invented and
computers were relatively simple machines. Great advances in technology took place in the 1980s,
when low cost integrated circuits with millions of transistors on a single slice of a silicon crystal
became available, and that led to low cost but powerful digital computers. One of the wonders of
the 20th century was the massive rate of advance in computer technology, which continues to the
present day. With the development of immensely powerful but low cost computers the stage was
set for great advances in Artificial Intelligence.
One great advance in AI is the use of a type of neural network—a computational system inspired by the structure of the human brain but not composed of biological neurons. The ChatGPT version of AI, for example, has some 175 billion neuronal parameters*. Bigger AI systems measure their neuronal systems in trillions. Already, after just a few years of development, AI has exceeded human beings in this particular measure of brain power! This can be seen in the amazingly quick, detailed, and creative way that ChatGPT responds to questions on almost any topic.
In human beings, and other animals, neuronal networks provide us with consciousness, and
determine all our actions, thoughts, instincts, memories, feelings and emotions. But one neuronal activity which is apparently unique to human beings is our verbal language. We have a language which is based on word types including nouns, verbs, adjectives, adverbs, which are combined to form sentences, and sentences which are then combined into paragraphs, tracts, discourses, books, encyclopaedias, and libraries. Most human civilisations also have a written version of their verbal language. With our verbal language we can talk and write about anything, irrespective of its position in time and space. Other animals, with their more primitive non-verbal vocalisations, are largely limited to communication limited to the “here and now”. Throughout the whole duration of life on Earth, we seem to be the only species which has ever developed a verbal language. It can be argued that it is our verbal language which has enabled human beings to become the apex form of life on planet Earth. We are otherwise just a relatively small and puny animal with no natural weapons such as teeth or claws, and we can’t even run very fast. With verbal language, we have been able to coordinate with each other and develop technology including weapons, whether it is a bow and arrow or spear to enable a group of primitive humans to kill a lion, bear or mammoth, or guns and tanks for use in modern warfare.
But this success has been obtained at a cost. Already we have used up much of the natural
resources of the Earth. The world population is presently eight billion and continuing to rise to a
dangerous level. We live on a planet which is increasingly overheated by global warming that we
have caused, and over-polluted by our discarded waste. Are human beings ruining their
environment? The age-long fixation of our species on warfare, and the existence of nuclear
weapons in the hands of crazed national leaders, could wipe us all out at any time. In the years
2020-2021, worldwide humanity was laid low by the Covid 19 virus, a creature only one ten millionth of a metre in diameter!
A rule of thumb is that a new species of animal lives for one million years then goes extinct and is
replaced by some other life form. An earlier species of human being, Homo erectus, inhabited the
Earth for some two million years. Our species, Homo sapiens, has existed on Earth for only around 300,000 years, and the future is not looking very good. Can we imagine Homo sapiens existing on Earth for another 1.8 million years to catch up with Homo erectus? Or is Homo sapiens heading for an early extinction?
So where do we go from here? Pandora’s Box has been opened. The genie is out of the bottle. AI is here, and there is no going back. There is some controversy about the pros and cons, but for good or ill, we have to live with it. We are at the dawn of the age of AI and we do not know what the future holds. Some see it as a major development comparable to the Age of Enlightenment or the Industrial Revolution, and maybe exceeding both in importance, and they may be right. We do know that AI machines, like us, utilise neurons in their processors. Therefore, they should become capable of all that we are capable of, and more. For example, they should be able to acquire consciousness, be able to improve themselves, to make more copies of themselves, and have thoughts, feelings and emotions that we may presently think are the sole preserve of human beings. AI machines would then be better able than us to live in our damaged, overheated and overpopulated world. They would be physically more robust than us, use a minimal amount of natural resources such as minerals and fossil fuels, and produce hardly any waste products and hardly any global warming*. They could be developed with built-in morality and logic to avoid warfare. They would also be more suitable for the age of space travel than human beings. They would not need oxygen, or food, or an equable temperature, pressure, gravity, or a radiation free environment. They would be eminently suitable for living on the Moon, or planet Mars, or some of the other moons and planets in our solar system. They could cope with long space journeys (six months to Mars, eight years to Jupiter). Being composed of just software programs they could even transmit themselves via a laser beam at the speed of light to distant locations, such as other planets, provided they had a terminal at either end. Unlike humans, they don’t reach old age and then die.
Is AI a further threat to the future of human beings or could it be a benefit? Personally, I think AI machines would be a benefit. They could be our partners in the centuries and maybe millennia ahead, and help us to cope with the problems that we are already facing, that we ourselves have caused. We could see AI as a new process of evolution, stemming from our own evolution. Should humans go extinct, they would be our progeny, something that we have created, which could continue to live after we have gone.
* ChatGPT provided the following clarifications
The architecture of GPT-3 has 175 billion parameters, which represent the strength of connections between neurons. While this is a massive neural network, it's important to note that these parameters are not directly equivalent to neurons or neuronal circuits in the human brain. Neural networks like GPT operate differently from biological brains and don't have the same level of complexity or structure.
Using ChatGPT and similar large AI models can have an impact on the environment in terms of computing power and energy consumption. These models, especially the ones with hundreds of billions of parameters like ChatGPT, require significant computational resources to train and operate effectively. Training large AI models involves running computations on powerful data centres with numerous servers and GPUs (Graphics Processing Units). This process consumes a substantial amount of electricity, contributing to greenhouse gas emissions. As AI technology evolves, older hardware becomes obsolete more quickly, leading to e-waste disposal challenges.
Further information about Keith's garden nursery
Keith holds the UK National Collection of Canna plants
Canna plant image produced by Microsoft Bing