“Change” is a familiar word. It’s been used in the English language for hundreds of years and we all know what it means. Although, nowadays, when we are talking about our advancements in the technological fields, we prefer the word ”progress” over ”change”.We’ve been technologically progressing at such a pace that we completely forgot about the animals we used to be, and we aspire to be the gods we once worshiped. Our progress up until the past century was slow, but we were always on the same path: a path of innovation. Every invention, discovery or paper until now led us through this technological route that, some think, approaches an asymptote, a technological singularity. We are yet to make, mankind’s greatest creation so far, an artificial sapient being. Now creating life might not seem like a mystery to us anymore, but creating intelligent, sentient life forms is both exciting and terrifying. But right now, what are the problems that we are facing in the progress of doing so? What impact will these machines have on the socio-economical point of view?Fortunately, it is not the first time we are confronted with similar problems. The English Industrial revolution in the years 1760 to sometime between 1820 and 1840 had similar issues. Technological revolutions did not have only positive consequences on workforce. The Industrial Revolution also led to unemployment and at the beginning lowered real wages. Although there was a slight improvement in the standard of living, over that period there was intensified exploitation, greater insecurity and increasing human misery. The increase in production and efficiency came with a huge class gap between the rich and the poor and brought unhappiness to most of the people.Like the steam engine during the First Industrial Revolution, the Information and Communications Technology (ICT) has completely changed the way society organizes its economic activity. While the 18th or 19th century’s machines replaced manufacturing, the new thinking machines have been increasingly capable of performing conceptual managerial and administrative functions and of coordinating the flow of productions from extraction of raw materials to the marketing and distribution of final goods and services. This new computers based automation has led to a major decline of the global labor force whether in the manufacturing sector or in the newly evolved service sector, which had been absorbing for more than forty years the job losses in the manufacturing sector. How can we learn from these mistakes and improve the economy with AI? How can we do so without over-looking aspects that may lead to a destabilization of economical factors?PRESENTGeneral Approach to Current Applications in AIThere is an idiom stating there are 3 things that do not change over time: death, taxes, and change itself. Recent statistics prove the idiom wrong,”change itself” is changing. Statistics show that technology is developing 10 times faster than expected1. People prefer using applications instead of receiving human-expert help. They rely on algorithms rather than hand-work. The manager of Mercedes Benz said in one of his recent interviews that their opponents are not other car-producing companies; but they are technological companies like Amazon, Google or Apple2.We see that software computer programs have started to disturb industry environment nowadays. People tend to try to enhance software algorithms that solve every-day problems rather than trying to come up with a possible solution. They expect intelligent machines to do so. Uber is only an application we use to travel short distances; the producers do not own any company cars, yet Uber is the most used taxi-company. The biggest hotel-chain company in the world, Airbnb, does not own any real estates, and yet, people trust it and use it because of their groundbreaking ideas, marketing and efficiency.Affection of AI: Driverless CarsA paper written in 2013 argues that industrial robots are becoming more flexible and far cheaper than their predecessors3. Instead of paying workers a monthly wage, charging a robot which can do the same amount of work considerably cheaper. These robots are improved to perform simple jobs for manufacturers in different fields or sectors. David Rotman mentioned in his research about some website of a Silicon Valley startup called Industrial Perception. There are videos published in 2013 about a robot designed to be used in warehouses for replacing boxes. The videos also create some sensations such as “Google’s driverless car suggest what automation might be able to accomplish someday soon. We have partly reached those days, we see how automation in car industry comparatively affects our life. Cars can park themselves without any control from their driver nowadays, and totally driverless cars are going to be presented to public in 2018. Total number of fatalities is increasing incredibly fast compared to how it was during 1900s4.This lead scientists to search for alternative ways to reduce accidents to a minimum range. Augmentation in usage of cars also increased air pollution. Negotiations started on how much humans use and waste fuel, and how much that changes our world. Social platforms and non-governmental environmental organizations protest against human damage to nature, and this build up a pressure on car companies. Thus this lead companies to try to find and use variations of cheap and virtually infinite fuel such as electricity or bioenergy, instead of more common choices for fuel like gasoline and petrol. After years of work, we will be introduced to this machines that detect the surroundings around themselves, estimate the possible outcomes of their actions, calculate the fastest and/or safest way to a direction and drives you wherever you want. The usage of this intelligent machines are expected to decrease the number of car crashes, and the number of cars that are being used now. They will reduce the time spent during driving, and thus reduce the atmospheric pollution cars causes.Even though it should not be overlooked that this revolution could cause serious raise on unemployed population over years as we will no longer need special driver, licenses, buses vice versa.IBM-Watson ProjectProducing intelligent machines that can keep their balance on 4 wheels might seem easier compared to building actual human-like robots. With that much equipment, it is hard for robots to remain balanced whilst doing their task. Walking, recognizing obstacles and surmount them, meanwhile trying to do their given job such as lifting a box or passing an object might be too complicated for robots to do. Scientists still need to figure out more reliable ways to make machines do their job. Nevertheless, humanity came far more closer with scientific projects on building up problem-solving intelligent machines that do not require any movement. Let us look at Watson, an intelligent machine represented by IBM that answers questions posed in natural language. IBM scientists developed a robot that answers as much questions as possible by storing its memory with thousands of documents. What separated Watson with other robots was that it is not connected to internet, and it gives answers vocally. Watson searches through its memory to answer a specific question5. This method is a possible solution to one of our current problems in AI. The machines we produce need internet or Bluetooth to function, and we do not have that kind of capability under our hands all the time especially in underdeveloped countries, even though we want to enable the usage of these machines world-wide.IBM carried forward its studies on Watson and successfully made it use its knowledge to make diagnoses, give treatment recommendations, and do small tasks such as measuring tension. These small tasks require small movements and technology is reaching to that point speedily. It is now a concern that: will this technology disrupt healthcare in the same way it has in other industries? It has reported that Watson can diagnose cancer 4 times better6. It is envisaged that local clinics which does small operations, blood tests etc. might not be needed if the technology of Watson will have implemented to a phone application. On the other hand we might still have these little local clinics, but with robots working and managing the facility.Watson can answer questions by searching through its memory, but what if we build a machine that decides itself without using any memory. “We can make machines to think so that they can do more than just following the instructions step by step. But these thinking features also need to be programmed in advance.” says Harjit Singh7. A simple example is solving Tower of Hanoi problem, or Rubik’s cube. We can write programs that calculates every other achievable solution to a problem as computers does not get tired like human brain does. Another example is playing games against computers. In 2016, a computer beat the best go player in the world. This happened 10 years before planned. It is projected that in 2030, computers are going to be smarter than humans2.Baby ProjectTo be able to use intelligence, knowledge is required; and knowledge is based on learning and experience. The intelligent machines we are producing nowadays does not have time dimension like humans do. They only obtain knowledge in our control; they do not age over years, so their functioning does not get old and they do not struggle with remembering; they only get experienced in specified fields we let them to retrieve. This is exactly why Nils J. Nilsson suggested to make machines capable of that kind of learning. He argues we should build up educable systems that can learn and taught to perform. His idea is referred as ”The Child Machine Project”8. Babies gain knowledge by steps. There are stages that should be taken. If we successfully implement these stages into an education system, we might be able to teach to the ”baby robots”, and let these robots improve their intelligence by themselves. If we really can manage to produce these baby robots, imagine what will happen in next decades. An agent with capability of learning and evolving itself might also be able to teach. We already have applications on our phones that help us gain knowledge, smart-boards to help educational systems, websites to reach anything we want to. Having robots that teaches to another robots or humans will have a huge influence not only in educational policies but also in many other industries. Ethical IssuesAllowing intelligent machines to take a huge part of our lives is a heated debate in our decade. Companies are looking for ways to gain most money with least employment. Unemployment caused by mechanization bothers various of business sectors. Even though people rely more on algorithms, it is a known fact that incontrovertible amount of people are actually afraid of intelligent systems. Scientists are building up machines that takes orders. But what happens if a robot is given two conflicting orders? Whom should it obey? Converting right from wrong is a hard task even for human beings. People go to courts to display the most correct judgment in various situations. There are lawyers, judges, policemen who tries to provide justice. It is an anticipated subject for AI world that how are we going to deal with such ethical issues. One right for someone might be wrong for another, so the robots9 choosing will depend on its producer. Technology is not always used by good-hands. The usage of technology is inspected by laws, but there are situations where it is not possible for scientists to maintain their control on the machines they work with as there are numerous of hackers that steals information to sell to dark web market. If we cannot fully control human-beings yet, how are we going to control intelligent machines which learns non-stop, gains knowledge and information, decides by its own, and never gets tired?10But although, right now, we are trying to answer the pressing questions of the present, where is this path going to lead? Can we build an reliable economical system that will apply for the future applications of Artificial Intelligence as well? What will these “future applications of AI” look like?FUTUREExpected effect on the job marketIt’s hard to talk about the future of AI without mentioning the elephant in the room: loss of jobs due to replacement of humans by intelligent systems. This is already happening at present day, mostly not with intelligent systems but with automated systems instead. Particularly jobs that are very repetitive such assembly line work and jobs that don’t require creative and social intelligence. According to a study done by C.B. Frey and M.A. Osborne, 47% of all jobs in the US is in a high-risk category, meaning there is a 70 to 90 percent chance that they will be replaced within 2 to 3 decades11, while 19% is in a medium risk category. These statistics are applicable to the first world as a whole, with experts’ opinion on the year higher-level machine intelligence will be reached (meaning intelligent systems will be better at every single job than humans, though this doesn’t mean more cost-effective) on average being the year 208112. This automation of high-risk jobs will cause a shift in employment from the high-risk category to the low-risk category, and since this category doesn’t contain enough jobs to sufficiently take care of all newly unemployed applicants, the unemployment rate will rise. This, in turn, will lead to a society where the people who used to work in high-risk sectors are now almost fully dependant on the people who work in low-risk sectors. Perhaps due to innovation, or due to reform of the current economic system as a whole. One such reform is the introduction of a basic income. In his book Basic Income: An Anthology of Contemporary Research K. Widerquist gives arguments for two types of basic income: one being a subsistence income, which would enable employment to become intermittent (less-than-full-time and non-regular work hours), or even to encourage this type of job13, and one being a sufficient social income. The subsistence-income-based system aims to reduce working hours required to acquire a relative wealth, while still maintaining the need for jobs. Because of the introduction of the basic income, people would have to work less hours in order to attain the income they currently have, causing more low-risk jobs to be available and thus helping the problem of intelligent systems taking over the job market. You could argue, however, that this is but a temporary solution since low-risk jobs still have a risk to be taken over by intelligent systems and considering the world population is continuously increasing, there would be a day once the low-risk jobs start running out where this system needs to be replaced by another system.In a sufficient-social-income-based system the aim is, according to Widerquist “not to force the recipients to accept any kind of work on any terms whatsoever, but to free them from the constraints of the labor market. The Basic Social Income must enable them to refuse work and reject “inhuman” working conditions. And it must be part of a social environment which enables all citizens to decide on an ongoing basis between the usevalue of their time and its exchange-value: that is to say between the “utilities” they can acquire by selling their working time and those they can “selfprovide” by using that time themselves.” (298).Limiting factors of AIFuture applications of artificial intelligence, it’s integration into goods and services used by society, isn’t so much about increasing the capabilities of artificial intelligence as it is about increasing the range of fields where it can be used, particularly fields where work that is nowadays done by hand will be done by an intelligent agent. This may sound very unspecific or broad and that is because it is. Any operation that requires human attention has the capability to be replaced by an intelligent agent once this agent is capable enough, which is only a matter of time. A real limiting factor for this application is not the AI itself, it is the cost-effectiveness of replacing a human worker with an intelligent agent, but the performance-factors of AI are largely undocumented14. Another limiting factor is the upscaling of solutions to certain problems in AI from a testing environment to a real-world environment15. Based on this given, the introduction of intelligent agents that can act on the world is likely going to happen in functions where humans aren’t particularly efficient.Examples of future applications of AI One such field is intrusion detection (the identification of attempted or ongoing attacks on a computer system or network), whose systems were originally built by hand and have already partially been replaced by an artificially intelligent system that constructs part of a system based on certain specifications, and it’s expected for it to continue taking over parts of building an intrusion detection system, particularly feature selection, customization and various applications that deal with data manipulation16. Monitoring this field may give an indication as to when integration of AI into functions originally done by humans will become more commonplace.For examples of integration of AI into society that stand closer to the average human, look no further than health care, particularly psychological practice and treatment of certain types of cancer.Radiation oncology has always been deeply rooted in medical modeling for prognostic and therapeutic purposes, and the introduction of machine learning to draw conclusions from models that model similar symptoms that a patient has is a promising future direction for this field. Currently, the standard for looking at the most likely outcome of a treatment for a patient is to do a clinical trial, but this is costly, slow, and requires multiple people to work on it. Replacement with a machine learning algorithm that can analyze countless models to see which one best fits the patient in order to draw conclusions such as which treatment to use will save a lot of work, and decrease the time it takes for a diagnosed patient to get treated. The same also applies to analyzing the symptoms of a patient to draw a conclusion of which illness a patient is suffering from4. One study of the accuracy of the predictions of a certain algorithm called Vector Machine-Based Prediction compared to the accuracy of a multivariate logistic model showed that the machine learning algorithm is, on average, roughly 8% (from 67.2±0.8% in the traditional method to 74.9±0.7% using the machine learning algorithm) more accurate than the multivariate logistic model which is currently widely used to predict what the best approach is to controlling the spread and growth of tumours17. The consequences of wide-scale application of machine learning algorithms in diagnosis and prediction of treatments for diseases would save a lot of expenses in the forms of using inefficient treatments and misdiagnoses as well as drastically reducing the time it takes to go from diagnosis to the start of the treatment.ConclusionThe CEO of strategy firm Webbmedia Group, Amy Webb, wrote: “There is a general concern that the robots are taking over. I disagree that our emerging technologies will permanently displace most of the workforce, though I’d argue that jobs will shift into other sectors. Now more than ever, an army of talented programmers is needed to help our technology advance. But we will still need folks to do packaging, assembly, sales, and outreach. The collar of the future is a hoodie.”We all understand the problems Artificial Intelligence raises from an economical and ethical point of view. Because, now,we only speak of AI as something that we, as humans, own. A product of our ingenuity. Since the beginning of human history, great civilizations have been built on the back of disposable workforce. From the ancient Greeks,to the European Colonial Empires. Maybe now, we are close to creating the perfect “slave”. A slave that the can do virtually anything, and maybe only be restricted by our own minds and imagination. Or maybe we are creating something bigger than ourselves, a force we should not mess with. It’s only a matter of time until we start seeing the other side of the coin when it comes to AI.You might have heard about Sophia. Sophia is a humanoid robot developed by Hong Kong-based company Hanson Robotics. It has been designed to respond to questions, and has been interviewed around the world. In October 2017, the robot became a Saudi Arabian citizen, the first robot to receive citizenship of any country (Wikipedia). Maybe right now, Sophia’s case regarding the citizenship might be more for political or marketing reasons, but it will just be a matter of time until we will start seeing the other side of the coin regarding humanoid robots. Then, it is highly probable that we will start reevaluating our ethics regarding the exploitation of AI related workforce.In conclusion, even though most of our Artificial Intelligence related issues are as pressing as they are hypothetical, they are important. We can only assume, and try to prepare for, the next big ground-breaking discovery and its outcomes. As time has showed us, our world is changing faster and faster, and we can only adapt to it the best we can.