JonathanRodriguez-RojasDaniel EstradaCourse: HSS40512/18/17Artificial Intelligence Will not Be Taking over HumansRecently, the media has been filled with speculations concerninghow Artificial Intelligence (AI) would in the near future take over all thetasks that humans can do. Even some very ingenious people including Bill Gates,Stephen Hawkings, and Elon Musk appear to have been swayed by the idea thatbecause the computer is increasingly becoming ever faster, they will surpasshuman intelligence at some point (Robinson n.p.).Some people are now terrified that intelligent machines will substitute thehuman race, as depicted in many fiction films like The Terminator.
Prominent technologists are already proposingreplacing the existing models of public service with algorithmic regulation.This proposition aims to subordinate decision and policy-making roles of humansto automated systems. This perspective not only cedes much control over thelives of people to AI but also exaggerates what technology can do (Robinson). This paper will explore thereasons why AI cannot replace humans based on the fact that decision-making andunderstanding involve more than just intelligence.
Moreover, the criteria thatcomputers must meet to ever equal the sophistications characteristics of human brainswill also be discussed. First, the misconception that AIwill take over humans is misinformed by Moore’s Law. This law predicts thatcomputing capacity will increase exponentially (Moore).According to Moore, the number of transistors that could be integrated into asilicon chip would double every two years with advances in developing denserchips evolved.
With the increase in processing power, the capacity forprocessing more information in complex forms have come to be compared with theprocessing capability of human brains. This comparison is wholly mistakenbecause processing power does not equate to intelligence, as intelligence doesnot corresponds to the ability to make objective decisions. Neither does thecapacity to decide based on information translates to the ability to judgethings on the basis of values. Most definitions of intelligence exclude the actof decision-making, the ability to choose objectives, or hold values thatinform the process of decision-making. Most of the disciplines of AI entailscomplex processing of information. Usually, the objective of that processing isto select answers of an action course from a range of alternatives or from amass of structured information, which allow robotics to execute some tasks suchas in self-driving cars, or computer applications that compose music. Thesefunctions, certainly, meet the contemporary definitions of intelligence.
On theother hand, the definitions of humans as living systems surpasses thisdefinition of intelligence. Humans are defined by the will and capacity tochoose the criteria and objectives according to the perception of values gainedfrom experiences besides the mere intelligence necessary to make decisionsaccording to the knowledge founded on specific criteria and objectives. Various old stories, suchas the “Liar!” have captured the incapacity of technology in dealingwith ethical dilemmas. The story highlighted the difficult involved in logicalanalysis to resolve the ethical problems that humans encounter on a regularbasis. Technology savvy experts are yet find an algorithm for dealing with thisdilemma in a manner consistent with values and behavior of humans. Yet, alltasks involve anticipating or furthering specific human values or behaviors. Note worthy, the most concerted efforts to solve the mentionshortcoming in logical systems were conducted in the twentieth century byLudwig Wittgenstein and Bertrand Russell (Cited in Robinson).
These expertsdeveloped the theory of Logical Atomism which posited that the whole worldcould be described regarding atomic facts. However, their diligence andconsistency failed to bring the theory to fruition. Logical Atomism failed toprove that the world could not be described in that manner.
In the same vein,the demise of the theory illustrates how it is impossible to apply logicalrules to relate facts between different levels of abstraction; for example,oxygen-carrying capacity of red blood cells to abstract that physical assaultis criminal. This metaphor implies that as humans increasingly create powerfulcomputer systems that use increasingly sophisticated logical systems, thesemachines may succeed in mimicking human thinking. However, such complex logicalsystems will always fall short in certain situations that require humansdeploying their judgments based on values that just humans can empathize withbecause they are often informed by experiences to which they can relate. Furthermore, a computer programs and data cannot fully describethe world and the objects that people care about. As the last greattransformation of the industrial revolution largely complemented the humancapacity of humans to produce, so would AI. The current hype about AI replacinghuman intelligence is just but a bubble. Computer technology cannot replacehuman workers contribution to the global economy but complement them.
Nevertheless, the economy of developed countries will betransformed within the next few decades. Indeed, automated systems will replacemany current jobs (Susskind R and Susskind R n.p.
), and workers will have tolearn new skills or remembers old skills in some situations to help themperform tasks that would uniquely reflect human attributes. Robotics or AI systems would never evolve to outperform andoverthrow humans. This scenario would always be an illusion so long as thesesystems continue to take decisions and actions, however sophisticatedly, toattain a defined outcome that humans themselves set. Humans will never set theobjective of its extinction (Robinson). For a technological entity to set suchan objective, it must have learned through experience to values its existence,a fact which is impossible. Although computers can choose based on the available data, this isnot the same as Judgment. Judgmentsare informed by values derived from experiences rather than availableinformation. Computers have yet to experience life as humans do, and do notdevelop values.
Similarly, they do not have a purposeful foundation forchoosing or setting goals, or objectives, as they behave based on theinstructions that humans give them. Thus, it is erroneous to regard thedecisions that logical systems make as judgments, and the role they can play inour lives and society are fundamentally limited. They would only be able toachieve this objective if they could experience the world in a manner thatwould allow them to develop senses of the value of their existence. Thus, theywould be able to achieve life-like behavior than many contemporary perspectiveson artificial intelligence. It is then that AI systems would make Judgments based on the values that theywould discover.
As of now, it is not even aware of its existence. Even iftechnology evolves to this level, they would often be limited by their susceptibilityto the environment. AI systems inability to attain free will constitutes their majordownside. So far, no known technology can exercise free will; their actions aremainly based on a program that we humans have set. Free will is an essentialcomponent in developing values, as it entails making individual objectives anddecisions from which comes the responsibility to handle their consequences. Kauffman explored these relationships in his paper “AnsweringDescartes: Beyond Turing” (Kauffman n.p.).
In the paper, he argued that any system that is based on logic or physicsconcepts cannot achieve free will because such systems are deterministic,meaning that what has already happened determines the next action or event. Basedon Kauffman, such systems do not provide an opportunity for a consciousdecision to influence the future that has not been shaped by the past. Robinson argues that even if more state-of-art technologies, suchas biotechnology, quantum computing, and nanotechnology achieve the creation ofhuman experience and behavior beyond the accomplishment so far of the digitallogic and classical physics, such development will be artificial life ratherthan AI. Artificial life models could potentially experience the world as wecurrently do.
If they become sufficiently sophisticated as us, their experiencecould result in free-will, values and judgment would be inevitable. However,those values developed by artificial life forms will not correspond to thenatural human values because there is no conviction that the judgments derivedfrom those developed values would be in our interest as humans. Companies specialized in creating robotics have made greatstrides. Their success, however, has occurred because they have effectivelycircumvented some hard obstacles to reproducing certain human capabilities,including common sense, free will, and value-informed judgment. They havefocused instead on improving the system’s sense of the physical environment,developing algorithms, and processing human language. These algorithms allowthe robotics to learn using feedback loops and self-adjustment. While thisprogress is remarkable, it is a far cry from replicating the major life-likeattributes characteristics of humans, including experience, values, and subsequentjudgment.
It is only when AI systems with these characteristics are replicatedthat humans risk to be overpowered and replaced because such computer systems wouldhave their values and their judgments would guide their actions, and anyconsideration of the human interest would come second to theirs. Even if the dreaded happens, AI intelligence would still not beable to replace humans in all fields of economy. One such economy is that ofhuman experience. People have very strong attachment to their experiences aboutan aesthetic or art, be it in the form of music, drawing, scenarios or picture.Digital technology cannot experience or even understand, and thus, cannotreplicate them (Robinson).
Moreover, the way digital systems process information cannot matchthat by which humans think (Torra, Narukawa andLong 67). This is best demonstrated by the chess contest between thecomputer program and Grandmaster Gary Kasparov, the then world champion forabout two decades, in which the machine beat the human champion (Metz).Nevertheless, despite the advances in computer technology that has beenachieved since then, a team of computers and people holds the position. Ofnote, this team boasts of neither the best player nor computer chess program inthe world (Robinson).
Rather, it has the best technique for disintegrating anddisseminating the thinking involved in playing chess between its human andcomputer members capitalizing on their competencies. This example bestdemonstrates how AI would complement us rather than replace us completely. The quality that we assign to experiences cannot be replaced bywhatever form of digital technology that would emerge. This is why so manyprogress in communication and media technology advances have not squashed theatrebusiness, as people still flock to theatres and concerts because of theemotional experiences they derive from watching live actors and bandsrespectively. No doubt, AI cannot reproduce different musicians and actors onthe stage. Experience economy is one area that human will continue todominate. It combines an aspect that technology cannot do on with what manypersons can do, including cook, and create among others (Pine and Gilmore 2).
If technology competes with humans onproducing and distributing certain products like music and movies, people willbegin to value of experiences the product at the expense of the value of thephysical product itself. For this reason, some artists give digital copies oftheir work for free but charge highly for personal performance in their fan’shomes. In conclusion, AI cannot completely replace humans from all thecurrent jobs they hold in the future with remarkable advances in technology. Thesemachines cannot mimic the traits that are essential in doing most of the jobsincluding common sense, free will, and values that inform their judgment anddecision. Although some breakthroughs in AI have been realized, they cannotmimic these essential human features. Therefore, humans should rest assuredthat advances in AI will never achieve full replication of these key humancharacteristics.
Works Cited Kauffman, Stuart. “Answering descartes: Beyondturing.” Cooper, Barry S and Andrew Hodges.
The Once and Future Turing: Computing the World. Cambridge UniversityPress, 2012.Metz, Cade. In a Huge Breakthrough Google AI Beats a TopPlayer at the Game of Go. Wired. https://www.wired.
com/2016/01/in-a-huge-breakthrough-googles-ai-beats-a-top-player-at-the-game-of-go/Moore, Gordon. “Cramming more components onto integratedcircuits.” Electronic Magazines2006.
Pine, Joseph and James Gilmore. The experience economy . Harvard Business Review Press , 2011.Robinson, Rick. 3 humanqualities digital technology can’t replace in the future economy: experience,values and judgement.
12 April 2015. 17 December 2017
Torra, Vincent, Yasuo Narukawa and Jun Long. Modeling decision for artificialintelligence: 8th International conference, MDAI 2011. Changsha, China :Springer , 2011.