With the technological explosion of data and the rise of social media platforms, the world is faced with new challenges coming from these technological developments. The changes in data collection and analysis are changing the way we do business and are having political and economic impacts that we have only begun to understand.
In this article, we will look at specific systems that utilize data and probe where these developments are leading us and what are the potential risks not just to the economy but perhaps the future of humanity. Specifically, we will be looking at the various aspects of AI, or artificial intelligence systems. The developments of these systems are in products we use every day, they are not just a future thing. Products using AI systems range from language translators to Siri, to autonomous vehicles and industrial robots. Our lives are being shaped and changed by exposure to machines.
Technological developments do change the way we conduct our business and personal lives, just look at the development of automobile transportation in the 20th century. The results were both beneficial and also caused hurt to humans and their existing business systems. The development of the internal combustion engine changed the look and feel of our lives. The horse-borne industries disappeared as automobiles were adapted and people lost their jobs. However, millions of new jobs were created in producing and fixing cars, providing the fuel and distribution systems to run them on a global level and roads had to be developed to make travel viable. Machines that learn can help humans solve problems but the path we are following is not without risks. We are learning to live with some of these risks already, but the future risks seem to be coming faster than we had previously imagined, and for those who are much younger than I, may become challenges they face in their lifetime.
Defining AI
Artificial intelligence has different meanings and widespread utilizations. One view is the created technology that allows computers and machines to function better; machines that replace human labor for mental work, with a goal towards more effective and speedier results. Others see it as a “system” with the ability to interpret external data, to learn from such data, and use those learnings to achieve specific goals and tasks through flexible adaption. Despite varying definitions, the common understanding of AI is that it is intelligence designed by humans and demonstrated by machines.
Because AI can be used in a wide range of industries, its uses are widespread and flexible and are impacting the industrial world in a wide range of applications. Manufacturers use AI systems to speed up manufacturing lines, test products, do research on new products, and plan transportation and logistical solutions to business problems. There are AI systems that can drive vehicles and fly aircraft without human drivers and pilots. The term AI is also used to describe some of these functions of human-made tools that emulate the “cognitive” abilities of the natural intelligence of human minds. Along with the development of cybernetic technology in recent years, AI is emerging in a lot of products we use today, and it will have a greater impact in the future. Think of optical character recognition, or the Siri, “speech interpretation and recognition interface” of the information searching equipment on your computer. These products are here and there is a wave of new products that will emerge in the future that will utilize these technologies and systems to a greater degree. The usage of AI systems can be very narrow, such as speeding up one manufacturing line, or may have wider applications, such as planning traffic patterns for major urban areas.
History of AI
Artificial beings with intelligence appeared as storytelling devices in antiquity and have been common in fiction. When access to digital computers became possible in the mid-1950s, one of AI’s first purposes was explored to utilize “step-by-step symbol manipulation research to explore the possibility that human intelligence could be reduced to “step-by-step” symbol manipulation, known as Symbol AI. These were the first steps to make machine learning global in scale. By the late 1950s, computers were learning checkers strategies, solving word problems in algebra, proving logical theorems, and speaking English. Researchers in the 1960s and the 1970s were convinced that symbolic approaches would eventually create a machine with artificial general intelligence and considered it to be their goal. However, in the early days of AI development, some of the early leaders of the industry failed to recognize the challenging degree of solving some of the first objectives and the cost of basic research. After a nice start and many promises, researchers ran into a financial wall and government funding was pulled. This led to what is termed as an “AI winter in the 1960s-and 1970s.
At the same time, the computer revolution was developing and a need for software systems to make things work was growing. The “space race” spawned a generation of smaller, more powerful semiconductors and the software to run these machines was developed to run a vast range of different applications. In the mid-1970s, there were few computers in American offices and mostly the “main-frame” types. By the early 1980s, personal computers were everywhere. At the same time, advances in the medical field were proceeding at a tremendous pace. The science-fiction dream of connecting humans to computers was developing on a realistic level. This promised advances for humans with medical problems but also raised some alarms of what a human-machine cross would lead to. A market need for faster methods of identifying and developing new medicines and new tools for industry was advancing.
In the 1980s, AI research was revived by the commercial success of expert systems, a form of AI that simulated the knowledge and analytical skills of human experts. Soft computing developed in the 1980s, such as neural networks, fuzzy systems, the Grey system theory, evolutionary computation, and many tools drawn from statistics or mathematical optimization grew at impressive speeds. AI gradually restored its reputation in the late 1990s and early 21st century by finding specific solutions to specific problems. Faster computers, algorithmic improvement, and access to large amounts of data enabled advances in machine learning and perception. Data hungry deep learning methods started to dominate accuracy benchmarks around 2012. In 2017, one in five companies reported they had incorporated AI in some offerings or processes. Some researchers have become concerned that AI was no longer pursuing the original goal of creating a versatile, fully intelligent machine. Most of the research now involves solving specific problems, even highly successful systems such as deep learning are being used for narrow purpose applications.
Systems of AI Research
The general problem of simulating (or creating) intelligence has been broken down into various sub-problems. These consist of the identification of particular traits, or characteristics. These are capabilities that researchers expect an intelligent system to display. The traits getting the most attention today, are reasoning and problem-solving. Early researchers developed algorithms that imitated the step-by=step reasoning humans use to solve puzzles or make logical deductions. By the late 1980s and early 1990s, AI research had developed methods for dealing with uncertain or incomplete information, using concepts from probability and economics. The problem is that humans rarely use the step-by-step deduction that early AI research could model to solve problems. Humans use fast, intuitive judgments to solve most of their problems.
The second field of research is what is termed “knowledge representation”. That consists of an ontology of symbols formally described so that software agents can recognize and interpret. AI research has developed tools to represent specific domains, such as objects, properties, categories, and relations between objects. This involves knowledge about events, states time, knowledge (that is assumptions humans assume are true, until told differently, and will remain true, although a combination of new or associated facts could lead to a change in opinion). The breadth of commonsense facts the average human has is enormous. Formal knowledge representations are used in content-based indexing and data retrieval.
An intelligent agent can help make planning easier to implement and can also make predictions about how actions could change it. In classic planning problems, the agent can assume that it is the only system acting in the world, allowing the agent to be certain of the consequences of his actions. However, if the agent is not the only actor, then it requires the agent to reason under uncertainty and continuously re-access its environment and adapt. Multi-agent planning uses the cooperation of many agents to achieve a given goal. Emergent behavior that is used by evolutionary algorithms and swarm intelligence is used in this area. Other major usages of AI programs include, natural language processing; systems that allow machines to read and understand human language. This area includes the acquisition of knowledge directly from human-written sources, such as newswire texts. Modern statistical techniques include frequencies (how often does one word appear near another, keyword spotting, or searching for a certain word to retrieve information, and transformer-based deep learning, which finds patterns in texts). Some of the usages of these systems include defense analysts, who use AI systems to search out certain published words and the frequency of these key statements to ascertain if some global political trouble will emerge in any particular geopolitical area. Some stock market analysts use such systems to study a particular industry. In a recent article that looked at the semiconductor industry, some stock market analysts using AI programs noticed that executives for the semiconductor industry were becoming more worried about supply problems by counting the use of certain words, the frequency of the use, and even the look in the executive’s eyes, for clues that they were getting more worried than they were letting on.
Other major fields of AI research include machine perception, or the ability to use inputs from sensors, such as cameras, microphones, wireless signals, and active lidar, sonar, radar, and tactile sensors to deduce aspects of the world. Applications include speech recognition, facial recognition, and object recognition. Computer vision is the ability to analyze visual input. AI is heavily used in robotics. Localization is how a robot knows its location and maps its environment. When given a small, static, and visible environment, AI robots have an easy time in their performance. In dynamic environments, such as a living body, there are greater challenges. Motion planning is the process of breaking down a movement that often involves compliant motion, a process where movement requires maintaining physical contact with an object. Robots can learn to perform more efficiently despite the process of friction and gear slippage.
Weak and Strong AI
From the functions and abilities provided by AI, we can distinguish two different types. The first is weak AI, also known as narrow AI, that is used to perform a narrow task, such as facial recognition, internet SIRI search, or a self-driving car. This continues to be the primary use of AI in our daily lives. Although weak AI is generally helpful to humans, there are some dangers. Because weak AI is programmed for a specific task, bad actors could disrupt or change the programming. This could cause disruptions in the electric grid, damage to nuclear power plants, cause malfunctioning of self-driving cars, or even shut down robots in factories. Pirates of the future will use digital boats to capture their booty and cause mayhem (as they already do today).
The long-term goal of many researchers is to create a strong AI, or artificial general intelligence (AGI), which is the speculative intelligence of a machine that can understand or learn any intelligent task humans can, thus assisting the human to unravel the confronted problem. While narrow AI may outperform humans on tasks like playing chess or solving equations, its effect is still weak to human activity. AGL, however, could outperform humans at nearly every cognitive task. A strong AI is a different perception of Artificial Intelligence so that it can be programmed to function as a human mind. The goal is to be intelligent in whatever it is commanded to attempt, even to have perception and cognitive capacities that are normally ascribed to humans. In the process of learning, an AI machine may eventually take off to a stage that no human has control of. With extra needed algorithms, the machine may choose to ignore its human creator commands.
Remember those weak AI systems above. Imagine that the pirate is not some young kid writing bad code, but an AI machine whose purpose is to disrupt systems and capture data and wealth. In the future, AI systems may fight against each other to protect our interconnected systems, as the large mass of humans stands to the side and watches.
Do humans need AI?
History tells us that humans are always looking for faster, more effective ways to complete their work and the pressure of further development also includes the pressure to look for additional new and better ways of doing things. We enjoy an easier and more leisurely life than our ancestors and this is because of the development of technology. Initially, Henry Ford started down his trail not to be a car manufacturer, but the motivation was that he did not enjoy back-breaking farm work and wanted a simple farm tractor to make work easier. Henry Ford did not invent the car alone. He was just one of the thousands that contributed to the development that led to mass production of automobiles and on the way, sparked other industries to help, such as oil and a distribution system to provide fuel in what developed to be a global supply chain system. The development of one particular type of technology can often lead to other developments in allied technologies.
We know that technology is also inherently dangerous. You do not have to look very far to see that the development of technology can lead to destruction and death. The development of atomic energy is a case in point, as it can benefit humans with a clean form of energy, or it has the potential to destroy the human race in a radioactive nightmare. In the last few hundred years, the development of weapons systems is to kill and destroy our fellow human enemies at a more effective pace. The development of a super-intelligent robot that could eliminate all humans just because they are not as smart as they are, is not an unfound fear. When you develop something smarter than you, there is a risk that it may destroy you. Analysts in the AI field and some political leaders suggest that we program our robots with a code of ethics. This sounds a little funny from a race of humans that might tell a robot not to kill but spend a lot of money and time to develop weapons to eliminate other humans. What would an intelligent robot think of that logic? Currently, over fifty countries are researching or developing battlefield robots. These robots can be bigger than us, human size, or as small as germs. Autonomous weapons systems could doom the human race if they get out of control.
The author Yuval Noah Harari has published several books, such as “Sapiens” and “Homo Deus,” where he outlines the history and development of modern humans and points to a not too rosy future. He points out that 70,000 years ago Homo Sapiens experienced a cognitive change that helped us develop language that made it possible for us to cooperate in large groups and eventually drove our less cooperative human groups, like Neanderthals into extinction. He sees a risk that we are already on the road that will result in being dominated by artificial intelligence. He points out that for a long time, intelligence and consciousness went together. Consciousness is the ability to feel things, like pain and pleasure, and love and hate. Intelligence is the ability to solve problems. Computers and artificial intelligence do not have consciousness, at least not yet. They just have intelligence. Robots solve problems in different ways than humans and the big difference is because we still have feelings. In science fiction, it is often assumed that as computers become more intelligent, they will inevitably gain consciousness. He says an intelligent robot may be a scary thing, but it may be more frightening is that they will be able to solve problems without any consciousness or feelings.
Dr. Harari says that machines are already gaining power over us. Banks use complex artificial intelligence algorithms to determine who qualifies for loans and global financial markets are moved by decisions made by machines analyzing huge amounts of data, in ways even their programmers don’t always understand. This is creating greater economic inequality. In the future, the countries and companies that control the most data will be the ones who control the world. Today in this world, data is worth much more than money. Ten years ago, companies paid billions of dollars for WhatsApp, for Instagram. And people wondered, “Are they crazy? Why pay billions to get this application that doesn't produce any money?” And of course, the reason was that they produced data.
Tensions of data and AI
The world is increasingly being divided into spheres of data collection. In the Cold War, we had the Iron Curtain. Now, it is the Silicon Curtain between the USA and China. Where the data goes, and who controls it, will be the dominant political and economic question of the future. Will it be Beijing or Washington D.C.?
Who owns the data will also be important. Who owns the medical data on your body, you, or some HMO? Currently, corporations and governments collect data about where you go, whom you meet, and what movies you watch. Very soon, trackers will have your health data. Soon, medicines will be tailor-fit to your chemical and biological make-up. Also soon, your medical bill will reflect the AI calculations on the probability that if you have a certain medical weakness, you will qualify for a higher rate. AI analysis can make for a better overall health system. However, if health insurance systems get your data, they will place a different probability on whether any intended treatment will work and certainly, tailers make estimations of your medical expenses. The AI analysis that selects your mortgage payments is already dividing people on an economic basis. Soon, AI systems will divide humans on a biological basis.
The future is both good and bad, but the downside risks are big
Digital life is helping develop human capabilities, As seen in the past, any big leap in technology is disrupting eons-old methods of doing human activities. Code-driven systems have spread to more than half the world’s inhabitants in ambient information and connectivity and have created unimagined opportunities. Artificial intelligent systems are helping us make complex decisions, helping us a reason and learn about situations, have helped us in speech recognition and language translations. AI systems make machines work across the globe seamlessly and save consumers and businesses time and money and offer individuals choices that offer a more customized future. The contributions to public health will continue and your medicine will likely become more personalized and effective. Over time, they may help develop clean, green cities and farms and have the potential to educate everybody. Artificial intelligent systems will be a plus for humanity for the most part.
On the other hand, there are big risks that we can’t sweep under the rug. The development of autonomous weapons systems could backfire on us. In recent years, we have seen AI systems use information as a weapon. Nation-states use robots to destabilize and weaken their opponents. The increase in lies and misinformation can dangerously destabilize human politics and undermine group efforts and traditional ties.
Our political system is already deeply polarized, and further divisions can be expected in a misinformation age. Aided by robots, outright lies, stories with little or no realistic backgrounds, deep fakes, and deliberate distractions will increase. Even in a free society, it is becoming more difficult to tell lies from the truth. In years past, we knew that candidates for office fib a little more when you get close to an election. In my youth, powerful interests influenced the outcome of elections by controlling the newspapers. Now, social media companies have a big hand in elections and lies and deceit are being directed by robot trolls. And it is difficult to know who is controlling those efforts. It could be our global advisories, somebody at home with their agenda, or merely AI systems that were implemented and not well understood. In truth, it is likely a combination of factors. Humans suffer from our hubris; we feel we can control things that may be outside of our control.
For today, enjoy the many benefits that AI is already beginning to bring to our world! But let me end on a less upbeat note. Technology is developing quickly to re-engineer our bodies and brains. Either with genetic engineering or directly connecting our brains with computers, or creating completely non-organic entities, we will be on the brink of creating a new human species. This new species is likely to be more intelligent than we currently are. The costs of new technology to move your consciousness into a cyborg body will favor the rich and politically important people, at least at first. The ramifications of this technology suggest a big dividing line between current humans and the new species that will emerge when this technology emerges. Aided by super-intelligent machines, the new species soon will be about as far from us as we are from apes. Homo Sapiens will split into two species. And you know what happens in history when a more intelligent and technologically advanced society meets a less intelligent one. The less intelligent one usually disappears.
