The development of full artificial intelligence could spell the end of the human race. Stephen Hawking
[…] we should be very careful about artificial intelligence. If I had to guess at what our biggest existential threat is, it’s probably that. Elon Musk
As this field continues to add new information, I thought I’d occasionally revisit the blog to update it on what is happening. In the past few days, Sophie, the Robot, came on view in the media. the article in Fortune is a good overview to present the rest of the blog’s issues. And, as I often reference cinema to complement my thoughts and Sophie is a bit spooky, and the earlier version below starts with a reference to cinema, here is another prequel reference from Ex Machina that is pertinent.
For the past two centuries, Western man was the first off the blocks in realizing the benefits of fledgling capitalism and the increased use of machines with inanimate power to accomplish what manual labor and tedious repetition had previously accomplished at a numbing cost and with poor efficiency. The premise of this musing came from recent articles that have added to the chorus of articles about Artificial Intelligence, which has been married to automated robotics to provide cheaper and more efficient labor. Those articles were in the Economist and the New Yorker, two vehicles of information dissemination that are themselves threatened by a form of Artificial Intelligence in today’s digitized, algorithmic world and are therefore important stake holders in the discussion. (there are many links at the end of this musing to many related articles)
But, what are the facets of this issue which need discussion? What information needs to be explored, who is controlling that information and what does that control entail and infer? As I often do when relating a topic to AI, it love to link it to cinema (my belief is that all education should value questions more than answers and that the most critical job of an educator is to present issues with many facets to explore…one facet to one person may outweigh another facet with someone else and the cost to society, in all measurements of cost, needs to be understood) . Cinema has been exploring the topic of artificial intelligence from almost the beginning of movie history. Of course, more recently the number of films focusing on this theme has increased dramatically as our own world is so influenced by robotics, algorithms and Artificial Intelligence.
Fiction has always dealt with societal issues and AI has been in its cross hairs from the first thoughts about the possibility of society employing AI and what that would mean to us. In fact, fiction is one of the areas that has benefitted greatly from constructing narrative with AI as a theme. Consider the long list of books and films that have AI as the central focus: Karel Čapek was the first to coin the term “robot” in the early 20th century, and Philip K. Dick, Isaac Asimov, Christopher Nolan, Ridley Scott and Denis Villeneuve more recently have explored the Utopian or Dystopian worlds AI can help create. But, there are other institutions in society that will eventually have to weigh in on whether AI is going to move us towards Utopia or Dystopia.
At this point one needs to explore the concept of ‘intelligence’, the second part of the AI initials. Humans possess a wonderful ability to reason, though we do not always utilize this fabulous tool evolution has provided us, often seeming to act without intelligence. Yet we now control its definition and employ computers that drive decisions and motions from algorithms we have written. That intelligence is using deductive or inductive logic applied to known or implied circumstances, the product of which is a reasoned prediction of what will happen in the future and what we therefore want that future to be. Man is the best animal at understanding his place in the world contextually, realizing self, time and hope. Also, we are capable of creativity and innovation like no other animal. What, though, I argue is even more critical in the human process of reasoning is to have curiosity. Curiosity leads to questioning, which tests hypotheses and leads to ever greater understanding. But, can one have too much?
There is an important factor these past two centuries that helped to define the eras within these centuries. This factor is the pronounced acceleration of the development of our ability to employ machines to do tedious, repetitive and heavy work. The steam engine was the first great assistant, followed by more meticulous machines that exploited its power to run mills, looms and such to replace the hundreds and thousands of human workers that had previously done this labor. In this first technology revolution much upheaval took place, leading to violent reactions, reforms and attempts by traditional individuals and institutions of power to protect their place in society. I mentioned in the last blog how important steel was to this phase of change, spearheaded by the likes of Andrew Carnegie, who became one of those wielders of power. His life overlapped the next era of change brought on by electricity.
Electricity continued this refinement into the 20th century of how labor was further enhanced or assisted in its goals. Then, later, globalization and improvements in transport affected who did the work, based on how the sum of labor costs and transportation affected competition and where the finished products were produced or sent. The shipping container, developed in 1955 by an American, brought on further geographical dispersement of labor and goods produced, which, with the fall of communism in general and the liberalization of its tenets in China in particular, altered the world irreversibly from 1990 onward. We are still adjusting to this change, the we being the common laborer in America in particular, and all the politicians in the world.
Throughout this time the idea of a robot, usually running on electricity, was introduced in the earlier 20th century, which also played out in the fictional ventures of cartoons, comic books, literature and film. Robots continually improved and were gradually ushered and interspersed into the labor force to build cars, furniture and such. Then, with the ideas of Alan Turing about machine computation developed in the Second World War, algorithms were integrated into the management of the machines that did repetitious work. Labor could be further controlled and manipulated within a unique environment, making it cheaper if you could mechanize and automate it.
Recently, digitization of information has allowed this manipulation to encroach on all manner of labor: manual, legal, the game of chess, assembly work, insurance actuary decisions, banking and finance calculations and predicting growth rates of economies, agricultural tilling and sowing, transport and vehicular guidance, and loading, unloading and stocking of product containers in ships, airplanes, trucks, trains, warehouses and stores. Automated machines, guided by AI, take over existing jobs traditional done by humans every day.
There are three major revolutions in industry that occurred over the past two hundred years, and we are poised for this fourth, which is already underway. The first, as mentioned, was spurred by the invention of the steam engine, followed by the use of electricity in nearly all manner of human endeavor, then the third was the information-technology phase which utilized digitization and algorithms. The one that is already happening and is hitting the world in different places differently, is where the economy is integrating robotics and artificial intelligence. This will eventually lead to what is known as the “dark factory.” In these factories all labor will take place through robotics run by AI. They will use less energy, cost less and function 24/7. Companies like Intel are already advertising and promoting the benefits of AI.
Without humans there is less need for lighting, for safety features, for hourly wages, for benefits paid out to employees, etc. When the first part of the Industrial Revolution occurred, with the advent of steam, laborers rebelled and destroyed machines in their Luddite reaction. Marx soon wrote about the impending doom to society brought on by the commoditization of labor, making it a simple, economic factor in the total cost of production. Ricardo predicted doom and gloom if one’s labor was falsely influenced. For him free trade and open competition led to the most efficient economies, even if the dislocation of labor occurred due to improved efficiencies. He was worried, though, that the downward pressure of wages would lead to instability. This did not happen as he opined, as laborers shifted to other jobs that were created for the most part. Yet, wealth did accumulate disparately, with the pie continuing to grow in size, but with the profits benefitting owners, bankers and lawyers disproportionately.
There was a reaction in many regions where some form of socialism was adopted. In these countries or regions, the economic model was regulated and legislatively overseen. Taxation was introduced to make conditions in society more equitable and to allow government to control the direction of the economy as much as possible and to reap the economic benefits of this growth for social projects. The harms that were perceived in those societies allowing the purest capitalism were given over to the purview of the government in many instances, who then passed laws regarding competition, to protect the environment and citizens’ health, and to attempt to ameliorate the volatility of the market. For more than a century during which capitalism grew and society changed drastically, this was a period of much confusion, disagreement and disparity. With only some corrections after the First World War and into the Depression, and later after the Second World War with the tremendous post-war growth, much of capitalism’s economic growth has been spasmodic, volatile and destructive to many, with a small percentage benefitting magnificently. Now we wonder if these conditions might be exacerbated by the use of Artificial Intelligence in manipulating robots to accompany our economy and to move it “forward”.
Now it is also possible for a computer and a camera to make medical diagnoses, or for a vehicle, trucks in particular, using cameras and thousands of computers to run autonomously. Walmart is using the same combination to assist manual laborers in restocking, cleaning and surveillance of its stores. Law enforcement is utilizing AI’s benefits to surveil, identify and assist in prosecution. AI has even been involved in aesthetic pursuits, determining in its limited way what the characteristics of ‘fine art’ and ‘genius’ are in works of art. Computers have even created works of art to rival human endeavors. A computer can even write a speech for a presidential candidate. Cameras armed with AI backup can identify species, uniqueness and patterns, all necessary elements in logical application of reason.
Therein lie the problems facing humans: How do we determine when this assistance is beneficial and when is it harmful? To ask this question in a meaningful way, we need to understand who is benefitting by AI’s implementation and who might be harmed, and, what is the nature of the harm done to anyone. How can we be protected? Who needs to be involved in that decision to determine if AI needs regulation, restriction or prohibition?
There have been presidential and private committees to explore this new era’s potentials. Each country is aware of the impending changes appearing or on the horizon and is either investing heavily in its implementation or is fumbling by fits and starts into the next stage of the world’s economy. Economics, a social science that is highly contentious in its positing of theories and solutions simply because it involves the human mind, is now faced with an immense problem. Many politicians are not ready to take on the problems they see on the horizon for a variety of reason: lack of knowledge, protection of a previous position or out of selfish, short=sighted aims. There will be winners and losers and who has control of the money is the most important factor in today’s world. That has not always protected those with the money and power in the past, when serious revolutions occurred to intercede in society’s control, and all estimates point to a even more contentious struggle as we navigate the cultural and economic waters of the next three or four decades throughout the world.
Keep in mind that the problem of AI is not the only problem facing economists and politicians. The Future of Humanity Institute also lists the threats to society found in nanotechnology, biotechnology, resource depletion and overpopulation. This does not take into account the vast problems already surfacing from climate change. If it’s true that the insect population is being debased, that sea level rise is accelerating, that extreme weather storms are going to be more prevalent, that regions of many countries will become less habitable or even inhabitable, then these will increase the pressure to cooperate between and within countries. Without accommodating those who are harmed by these changes, societies will start to fray.
Where Marx was off the mark in his predictions, he was not far off on the descriptions of the harm capitalism could do if unchecked. As mentioned in earlier blogs, Adam Smith, too, was aware of capitalism’s dangers and hoped that the Christian-moral driven, invisible hand of the market would contain destructive abuse. For the past thirty or forty years there has been an enormous production of wealth in the world. But, it has not been distributed evenly or how many would claim as being done fairly. Is it inevitable that we are going to have individuals like Andrew Carnegie, Steve Jobs, Bill Gates and the Kock brothers making our decisions for us? This tension has entered into many countries’ politics, with the most egregious example being the United States in 2016. Some thoughtful people, reading the figures they have at their disposal from the past half century, argue that a major shift needs to occur. Some have even called for a basic income to be distributed by the government to those impacted by these changes. It is not assured that, like in the previous three technical revolutions when those who lost jobs eventually migrated to some other land or profession to survive, that there will be new, equitable jobs to replace those lost. Even McDonalds and Walmart are mechanizing their labor forces more over the next decade. Will we willingly go into the Golden Arches or drive up to its takeout window to be “served” from cooking, to paying, to serving, to clean-up without a human being ever having been involved?
An open letter has been crafted by The Future of Life Institute to address the issues facing the world as it incorporates AI. The Future of Humanity Institute at the University of Oxford has a long list of threats besides AI that threaten our society. What then are the recommendations and what information are they using to arrive at their conclusions? We have much to discuss and decide going forward. I hope we’re up to it.
Related articles and links follow:
Article from the above link… The centerpiece of Mr Piketty’s analysis is the ratio of an economy’s capital (or equivalently, its wealth) to its annual output. From 1700 until the first world war, the stock of wealth in Western Europe hovered at around 700% of national income. Over time the composition of wealth changed; agricultural land declined in importance while industrial capital—factories, machinery and intellectual property—gained prominence. Yet wealth held steady at a high level (see chart, first panel).
Pre-1914 economies were very unequal. In 1910 the top 10% of European households controlled almost 90% of all wealth. The flow of rents and dividends from capital contributed to high inequality of income; the top 10% captured more than 45% of all income. Mr Piketty’s work suggests there was little sign of any natural decline in inequality on the outbreak of the first world war.
The wars and depressions between 1914 and 1950 dragged the wealthy back to earth. Wars brought physical destruction of capital, nationalization, taxation and inflation, while the Great Depression destroyed fortunes through capital losses and bankruptcy. Yet capital has been rebuilt, and the owners of capital have prospered once more. From the 1970s the ratio of wealth to income has grown along with income inequality, and levels of wealth concentration are approaching those of the pre-war era.
Mr Piketty describes these trends through what he calls two “fundamental laws of capitalism”. The first explains variations in capital’s share of income (as opposed to the share going to wages). It is a simple accounting identity: at all times, capital’s share is equal to the rate of return on capital multiplied by the total stock of wealth as a share of GDP. The rate of return is the sum of all income flowing to capital—rents, dividends and profits—as a percentage of the value of all capital.
The second law is more a rough rule of thumb: over long periods and under the right circumstances the stock of capital, as a percentage of national income, should approach the ratio of the national-savings rate to the economic growth rate. With a savings rate of 8% (roughly that of the American economy) and GDP growth of 2%, wealth should rise to 400% of annual output, for example, while a drop in long-run growth to 1% would push up expected wealth to 800% of GDP. Whether this is a “law” or not, the important point is that a lower growth rate is conducive to higher concentrations of wealth.
In Mr Piketty’s narrative, rapid growth—from large productivity gains or a growing population—is a force for economic convergence. Prior wealth casts less of an economic and political shadow over the new income generated each year. And population growth is a critical component of economic growth, accounting for about half of average global GDP growth between 1700 and 2012. America’s breakneck population and GDP growth in the 19th century eroded the power of old fortunes while throwing up a steady supply of new ones.
Tumbling rates of population growth are pushing wealth concentrations back toward Victorian levels, in Mr Piketty’s estimation. The ratio of wealth to income is highest among demographically challenged economies such as Italy and Japan (although both countries have managed to mitigate inequality through redistributive taxes and transfers). Interestingly, Mr Piketty reckons this world, in which the return to capital is persistently higher than growth, is the more “normal” state. In that case, wealth piles up faster than growth in output or incomes. The mid-20th century, when wealth compression combined with extraordinary growth to generate an egalitarian interregnum, was the exception.
Sustained rates of return above the rate of growth may sound unrealistic. The more capital there is, the lower the return should be: the millionth industrial robot adds less to production than the hundredth. Yet somewhat surprisingly, the rate of return on capital is remarkably constant over long periods (see chart, second panel). Technology is partly responsible. Innovation, and growth in output per person, creates investment opportunities even when shrinking populations reduce GDP growth to near zero.
New technology can also make it easier to substitute machines for human workers. That allows capital to gobble up a larger share of national income, raising its return. Amid a new burst of automation, wealth concentrations and inequality could reach unprecedented heights, putting a modern twist on a very 19th- century problem.