Chapter 4 Believe in the Need for Us to think and become a benevolent Link in the Future of Evolution

Many if not all living beings are able to seek out light, which benefits them in one or more ways. The importance of light to vegetation and consequently to all forms of life has already been considered in Chapter

  1. Many forms of life can be shown to be capable of programming themselves to a work and rest routine based on periodic availability of light. In the case of human beings also, it can be shown that such rhythms operate seemingly without the aid of the human minds. Perceptibly, humans following their minds and senses excessively against circadian rhythms seem to come proportionately to grief. Such voluntary tendency to ignore natural/instinctive rhythms results from some aspects of the highly evolved intelligence of the human species. Are intelligent machines made by men also evolving similarly? Is the man to machine relationship healthy enough to man? And as the relationship progresses in leaps and bounds, what should be the norms?

There is no universally accepted definition of intelligence, as this Wikipedia entry points out. https://en.wikipedia.org/wiki/Evolution\of\human\_intelligence

One definition is "the ability to reason, plan, solve problems, think abstractly, comprehend ideas and language, and learn.” Scientists have attempted to trace the evolution of hominid intelligence over its course for the past 7 to10 million years and to attribute it to specific environmental challenges. Some scientists understand this evolution as a necessary process, but they also hold that it would be a great misunderstanding to see it as one directed to a given outcome.  Has intelligence been merely a survival accessory, like the venom of snakes or the speed of some wild cats far exceeding that of the fastest antelopes? Scientists agree that the acquisition of intelligence is the only adaptation which could have allowed the human species to establish complete domination over the rest of the natural world. Could our ape ancestor have indeed felt this kind of need, and craved for it intensely? Has the human species yet acquired enough intelligence to manage this responsibility? It is obvious that some large apes which are like us have not chosen to move away from their habitats and have not been environmentally challenged to evolve into more ‘intelligent’ apes

The manner of evolution of the human brain has been explained by John Hawks, a professor of anthropology at the University of Wisconsin–Madison, while answering a question by a reader of Scientific American, Emma Schachner of Salt Lake City: “Humans are known for sporting big brains. On average, the size of primates' brains is nearly double what is expected for mammals of the same body size. Across nearly seven million years, the human brain has tripled in size, with most of this growth occurring in the past two million years.”

Different approaches to looking at early skulls have given us evidence about the volumes of ancient brains and some details about the relative sizes of major cerebral areas. “For the first two thirds of our history, the size of our ancestors' brains was within the range of those of other apes living today. The species of the famous Lucy fossil, Australopithecus afarensis, had skulls with internal volumes of between 400 and 550 ml, whereas chimpanzee skulls hold around 400 ml and gorillas between 500 and 700 ml. During this time, Australopithecine brains started to show subtle changes in structure and shape as compared with apes.” For instance, the neocortex had begun to expand, reorganizing its functions away from visual processing toward functions of other regions of the brain.

Nearly 1.9 million years ago, a modest increase in brain size occurred, including an expansion of a language-connected part of the frontal lobe called Broca's area. The first fossil skulls of Homo erectus, 1.8 million years ago, had brains averaging a bit larger than 600 ml. Eventually the size rose to more than 1,000 ml by 500,000 years ago. “Early Homo sapiens had brains within the range of people today, averaging 1,200 ml or more. As our cultural and linguistic complexity, dietary needs and technological prowess took a significant leap forward at this stage, our brains grew to accommodate the changes. The shape changes we see accentuate the regions related to depth of planning, communication, problem solving and other more advanced cognitive functions.”

In the foregoing paragraphs, science has been shown to have handed to us these two significant conclusions:

  • Growth of our intelligence is reflected in the growth of our brain size to nearly triple that of our ape ancestor of about 7 to 10 million years ago.

  • Advanced cognitive functions are unique among mammals, and even among primates, to human beings, and they involve the human mind, which must be a pre-eminent function of the brain, especially of the extra portions of brain that accrued to man during his evolution involving growth of intelligence.

But have scientists established that the human mind is indeed entirely a function of the human brain? If it is, does a human brain function in part as a mind on its own, unassisted? The analogy intended here is to man made machines such as yesterday’s computing machines and tomorrow’s fantastic improvements in them.

In his bestselling book, ‘The believing Brain’, (a Robinson book, published in 2011 in the USA by Times Books, a Division of Henry Holt and Company LLC) the Southern California based psychologist cum science-historian Michael Shermer gives his considered opinion that beliefs come first, and explanations for the beliefs follow and help to reinforce them. He explains how the brain has a mechanism, involving the introduction of a very pleasing chemical substance such as dopamine into the tiny synapses between neurons whenever a pleasing thought enters a neuron, which helps that neuron to communicate with neighbours about any thought or belief which one neuron found agreeable. This creates a demand for more of the same and the brain creates all the thinking support necessary to get deeper into whatever it started believing since every such reinforcing step is pleasurable to it. Shermer says that our brains find patterns in the world around us and infuse them with meaning until they become beliefs. From then on, “we seek out confirmatory evidence to support those beliefs, reinforcing them”.  Shermer cites examples in religious and political ideologies, ideas of conspiracy theory and supernatural phenomena (namely phenomena defying rational explanations). People like him, whose thinking had led them to become sceptics at an important stage in their life, have been able to find or think up plenty of evidence supporting their scepticism. Many of their friends who similarly decided at some stage that they were believers, have argued themselves firmly into staying as believers. Human civilization has settled into perpetual camps in conflict, both in religious and political philosophy and practice.

Perhaps we have reached a situation when some of us must reanalyse the stands we decide to get stuck in so that we and the rest of the living world can evolve in the right directions. Man’s new-found ability to produce machines which can increasingly take over most of his functions effectively ought to bring into focus the danger of wrong choices.

The review of a new book now available from the MIT press through Amazon online reads as follows:

“How deep learning-from Google Translate to driverless cars to personal cognitive assistants-is changing our lives and transforming every sector of the economy. The deep learning revolution has brought us driverless cars, the greatly improved Google Translate, fluent conversations with Siri and Alexa, and enormous profits from automated trading on the New York Stock Exchange. Deep learning networks can play poker better than professional poker players and defeat a world champion at Go. In this book, Terrence J. Sejnowski explains how deep learning went from being an arcane academic field to a disruptive technology in the information economy. Sejnowski played an important role in the founding of deep learning, as one of a small group of researchers in the 1980s who challenged the prevailing logic-and-symbol based version of AI. The new version of AI that Sejnowski and others developed, which became deep learning, is fuelled instead by data. Deep networks learn from data in the same way that babies experience the world, starting with fresh eyes and gradually acquiring the skills needed to navigate novel environments. Learning algorithms extract information from raw data; information can be used to create knowledge; knowledge underlies understanding; understanding leads to wisdom. Someday a driverless car will know the road better than you do and drive with more skill; a deep learning network will diagnose your illness; a personal cognitive assistant will augment your puny human brain. It took nature many millions of years to evolve human intelligence; AI is on a trajectory measured in decades. Sejnowski prepares us for a deep learning future.”

Angela Chen interviewed Sejnowski just about a week ago about the book for ‘The Verge’ and was able to get clear on a lot of hype about AI, machine learning and deep learning. She also got a very good perspective on the extraordinary prospects being held out by deep learning. In olden days human labour was inexpensive, while computational output was both slow and expensive. We have already reached a stage where the situation is very different. Human operation has become very expensive. Computational efficiency has gone up severalfold and has also become less expensive.

The trend is irreversible. It can irrigate many aspirations with hope. Dependence on this new technology will only make human physical and mental effort less and less available from the majority, while extraordinary and pathbreaking intellectual work will be conducted by the new machines and their very elite, very thinly populated set of human guides. We can easily recall the stages that humanity has gone through since the middle of the previous century and realize how rapidly things are changing and how this change can continue to shrink the percentage of the new knowledge-aristocracy and how it will have to meet new challenges.

By knowledge aristocracy I mean the scientists and technologists who know independently of any others how to create, manage and control the most updated technologies in any field. Each such person will have the responsibility to keep in a position of optimal satisfaction, thousands, millions of others who would only be obeying the rules and following the procedures laid out by the knowing. A person who is now in the seventies who can still do arithmetic mentally (and does not have to, of course) and still write his texts without errors (and once again does not have to) will understand what I mean. There will always be pockets of this planet, where the main power grid can fail for some measurable interval. You can guess what that would mean for the affected packages of humanity who have unlearnt most of their natural faculties and could only go deeper and further in that direction, when they get more and more deeply-learned-machine help. You only need to look at the reducing requirements of any kind of human assistance for progress, from deliberations in the developed world for four-working-day-weeks.

That means that the machine now has a larger proportion of humanlike neural network. This proportion will keep increasing without the need to do a lot of painstaking new ‘connection’. The more human like neural network that an AI machine develops, especially through its own learning, the greater is the possibility that it can mix truths with fantasies and that it can either benefit or deceive its employer. In the interview with Angela Chen, Sejnowski refers to a significant event that occurred about six years ago. “December 2012 at the NIPS meeting, which is the biggest AI conference. There, computer scientist] [Geoff Hinton and two of his graduate students showed you could take a very large dataset called ImageNet, with 10,000 categories and 10 million images, and reduce the classification error by 20 percent using deep learning! Traditionally on that dataset, error decreases by less than 1 percent in one year. In one year, 20 years of research was bypassed because of the ImageNet. That really opened the floodgates.” As Scientists know more and more about the architecture of the human brain, all or most of the information will be added in the work on AI leading to closer convergence of neurology and AI research.

Scientists and technologists and the management experts working with them can always hope to have the last word to see that the superhuman deep learners will always be under our control. But under whose control? Even at the present level of yesterday’s technological advancement, while commercial greed takes care to see that a very large proportion of humanity has access to the latest version of most sophisticated devices, how efficiently is the average customer served with reference to its servicing? How much help does the average customer get in disposing the older version and its accessories without polluting his own or someone else’s environment? Is there real accountability now for the increase in health problems of the lay person who is obliged to use the gadgets that the market place thrusts on him from time to time? Modern life also makes real healthy living more and more of a dream. Forests are continuously encroached upon. A great variety of processed food is made available, taking away the natural inclination to use fresh green materials from Nature’s laps and the pleasure of preparing a part of it at home with one’s own hands. People will have more holidays and travel more across time zones and disturb their circadian rhythms no end.

I mentioned just a while ago the definite possibility of fake outputs from super intelligent AI help. Mankind has been so far taught that they can expect predictable outputs from machines. Imagine the amount of chaos that will reign everywhere when you will need special recruiting expertise not unlike HR managers to purchase moderately truthful robots!

Let us go back to the meaning of evolution itself. The cheetah wanted to have a reasonable chance of overtaking a deer in its chase. The deer wanted to escape most of the wild cats which could be hiding anywhere in its path. Their speeds increased. The chameleon wanted to escape from its predators and thought that the chances would be more if it could change its colour to green and back at will. It developed that ability. A lot of small animals a long while ago dreaded the prospect of being wiped out by the sheer movement of mammoths and dinosaurs among them and hid into their burrows perhaps wondering why the huge monsters do not get destroyed by a natural calamity. Some of them survived when a huge meteorite impacted the earth. None of the dinosaurs and their mammoth cousins did. The number of interesting faculties that almost every living, surviving species has developed over the years is indeed of the stuff of fairy tales. The prehumen ape wanted to roam at will on land and did not like being tied to the treetops. It worried about its safety away from the treetop. It eventually lost its tail and started walking on land, with a brain that became more intelligent. All these changes took thousands and thousands of years. Legends of long ago including those of tribal origin in different parts of the world talk about a deluge when a certain number of living beings were rescued miraculously by a boat. In some cases, the boat was pulled to safety by a huge fish. I simplistically suggest that behind all these changes and narratives there was the wish. Just as modern man wished to travel faster, to swim across water, to fly in the air, and science, through his intelligent brain helped him. Having worked hard to achieve his wants, modern man now wants to work less and less to achieve the same ends. So, his machines and he himself are becoming super-thinkers and super-performers. The time required for changes has also been getting shorter and shorter. This ape is developing faculties which enable him to find solutions to many of his problems faster. The successful sustaining of Moore’s Law () that requires doubling of the number of transistors in an integrated circuit every year or two for nearly five decades has been possible because of path-breaking inventions all along the way by rising human ingenuity in a short time, almost on order.

If the availability or otherwise of light can be communicated to very rudimentary vegetation, if the wish of a chameleon or an antelope can be communicated to its environment and result in the wish being granted, if positive thinking enables a sufferer to get past his suffering, if a placebo or some nominal medication can get a person freed of a sickness, if pleasing beliefs are reinforced in the brain through the reward of release of dopamine whenever there is a repetition, if brains of scientists in chosen fields get more and more productive , how closely indeed does Mother Nature seem to interact with all living things!  Especially with respect to intense wishes! Some would call them prayers, which are being answered! In any case, our readers will agree that the seeking intensity of scientists is really of a high order.

As must have been the intensity of our sages who were seeking to know their selves. In such seeking and finding, they also managed, when the contemplation went very deep in polarised focus, to find a universal self too, which they termed param brahma. Some of them could identify their selves totally with the universal self, and others too felt there was a close kinship. The sages also had something to say about intelligence, which they called buddhi, whose flowering into prajna on occasion turned out to be most enjoyable. Such flowering seemed to occur especially during their yogic practices of the meditative kind. As the highly inquisitive and persevering young scholar Nachiketas found during his dialogue with Yama, Kathopanishad, 3.3, one’s self seems to ride in one’s body which is like a chariot, with the senses attached to it behaving like horses which simply pursue objects that attract them. The Self’s mind functions as the rein that controls their course which is finally determined by Intellect/Intelligence, namely buddhi functioning as the charioteer.

In the process of deep learning, because of massive data input and the possibility of quantitative rather than discriminative absorption of data, the resulting intelligence without tested values is bound to adversely influence the beneficial effects of output. Intelligent leaders of society, especially from the academia, who understand the need for channelling new inventions entirely for benefitting mankind with next to no harm to the already miserably treated environment and with security for all users must oversee applications to avoid dangerous, purely commercial decisions. We can ill afford to unleash Frankensteins in our midst.

We have managed to evolve rather well. We can only be better if we prioritize unmixed social benefit and safety while developing deep learning further, in this interesting stage in the evolution of machine-man interface. Such responsible leadership can be insisted upon by all right-thinking persons, whether one does or does not accept the sanatanic concept of universal consciousness or the possibility of consciousness entering the picture in the activity of the human brain.