This research seeks to explore the benefits and threats of artificial intelligence (AI) and machine learning. A deep literature review is carried out bring to light the current status of AI. AI is defined, and the two main types of AI discussed. The myths and facts about AI are discussed to clear the misconceptions surrounding AI. Technological singularity is then explored to show the serious concerns raised by the development and the use of AI. An analysis of the literature is carried out to bring together the facts and views obtained from different studies. Future expectations of AI is also discussed bringing to light the opinions of singularity theorists.
The topic of artificial intelligence is approached with uncertainty and confusion due to the broadness of AI. Technology advancement continues to disrupt people’s lives creating good and bad outcomes. There has been a rush to maximize the benefits of new technology and at the same time taking cover against the possible negative outcomes of technology. Businesses and learning institutions have been at the forefront in funding research to unveil new AI capacities. As a result, application of AI has increased, and this is making lives easier as well as increasing business profitability. The gap between machines and human beings continues to close, and there are fears that machines will at one time replace human beings thus causing an ‘existential threat’ that is a common area of study among AI scholars. Machines that have learning capabilities like Google home, Alexa and Siri are becoming increasingly popular. The second half of the twentieth century was characterized by breakthroughs in information communication technology that caused major changes in knowledge management. The twenty first century is experiencing even more technological explosions that are raising the question of the relationships between man and learning machines. In 1965, Irvin john Good predicted that AI would bring about a knowledge explosion (Brooks 1991). His prediction seems to come to pass considering the massive application of AI and machines learning in our day-to-day lives.
Artificial intelligence (AI) is a computer science branch that focuses on the creation of machines that simulate the human thinking processes (Bench-Capon & Dunne 2007). Over the years, computer scientists have been striving to bridge the gap between the human and machine thinking process. They continue to make improvements on AI that brings machines closer and closer to human beings. The opposite of AI is natural intelligence that is found in human and animals. The scope of AI is unclear since the technology has found vast applications in different fields. For instance, in the military, it is used for simulation during training. Soldiers are immersed in a virtual war environment, and this makes them harness their fighting schools. In management, AI is used in planning and decision making. Decision support systems are common in management, and they use AI to analyze facts thus supporting managers when making decisions. Expert systems have also gained immense popularity in management. These systems simulate an expert in a specific field thus enabling the organization to access the services of an expert without having to hire a natural expert. AI was established as a field in academics in 1956, and this move was influenced by high optimism in machine learning and successful attempts to create machines that can create knowledge (Bench-Capon & Dunne 2007). Since then, many studies and experiments have been conducted leading to major improvements in machine learning.
Artificial intelligence is a broad field, and its application to different aspects of life continues to increase. There are two types of artificial intelligence namely strong AI and weak AI. There is no clear line separating the two types, and this makes it difficult to identify the level at which a learning device becomes a strong AI (Jacquette 1993). Computer scientists seek to develop machines that simulate human beings. The number of tasks that an AI system can execute is one of the major distinguishing factors between a weak and a strong AI. A weak AI system only performs one or a small number of tasks. On the other hand, a strong AI performs numerous tasks and is closer to attaining the features of a human though process. Understanding the two types of AI requires one to draw a continuum where on one side is the weak AI while on the other end is a strong AI. As the number of tasks a learning machine can do increases, it moves closer to the strong AI point. Most of AI systems in use across sectors are the weak AI as they perform specific tasks. Strong AI also referred to as true AI is still a work in progress since computer scientists have not yet come up with machines that can think like human and execute multiple tasks as human beings do. Some computer scientists think that true AI will be in existence between the year 2030 and 2045 (Chrisley 2003). These systems will have the capabilities to communicate, plan, make decision, reason, and solve puzzles just like human beings. Another point of distinguishing weak and strong AI is the development process. Strong AI systems go through development stages like human beings, and they are organic. On the other hand, weak AI systems are more static as their development when in operation is minimal.
AI remains one of the world’s hottest topics due to lack of a good understanding regarding the capacities and ramifications of this technology. Whenever people cannot understand a concept fully, they develop myths in pursuit to explaining what is beyond their comprehension. The area of AI has also led to objective studies that have brought to light specific facts about learning machines. It is always important to clarify the myths and facts with the aim of enabling people to have a deeper understanding of AI systems and their consequences on different aspects of life.
AI is a Threat to All Jobs –The relationship between employment and AI has been a contentious area. Some people think that AI will replace all professionals in the work environment since machines can simulate the human thought process. Computer scientists have come up with complex machines that are replacing people in production processes. Producers prefer machines to human beings due to some benefits accrued from process automation. Firstly, machines are consistent as they are not influenced by fatigue, sickness, or emotions. On the other hand workers’ performance is subject to fluctuations since they are affected by these factors. Secondly, learning machines are less costly to maintain compared to workers. The only major AI cost is associated with acquisition compared with employees who need periodic compensation. Lastly, using AI leads to flexibility as it is easy to replicate their use as compared to the more rigid workforce that needs hiring and training to adapt to changes in the work environment(Lawless, Mittu, Russell & Sofge 2017). Although it is clear learning machines have inherent advantages over workers, replacing all workers in the work environment is not practical. It is worth noting that automating professional services is difficult and this protects professionals from AI. Professional services require puzzle solving, planning, and analyzing factors that are beyond machine abilities. It implies that automating professional services would lead to poor quality services. However, the threat of AI to some jobs cannot be dismissed. Less skilled jobs are under threat since commercial robot manufacturers have come up with systems that can perform most in not all of the general tasks in the work environment (Lawless, Mittu, Russell & Sofge 2017). As a result, unskilled and semi-skilled workers end up being unemployed.
AI will Surpass Human Intelligence-The argument that AI will outdo human intelligence in future is a major conception. Human intelligence and AI have substantial differences and comparing them cannot lead to an objective assessment. Human intelligence is only comparable to animal intelligence since they two have similar characteristics thus allowing the utilization of scale assessment. When using a scale of one to ten to compare animals and human being, animals would rank closer to zero while human’s score would be close to ten. Human beings are better than machines in some aspects of intelligence such as planning, strategic thinking, and emotional intelligence. It is impossible to create learning machines that are close to human capabilities regarding these aspects let alone outdoing human beings. On the other hand, AI is better than human intelligence in some aspects such as recall capacity and computation speed (Lawless, Mittu, Russell & Sofge 2017). In other words, machines have already surpassed human beings in these aspects.
AI Will Lead to the Extinction of Mankind –Different schools of thought suggest that in future robots will take over the world putting an end to the human race. This view stems from the belief that AI will lead to the creation of complex cybernetic organisms that will defy their creator and start harming the mankind. Others suggest that people of ill will design harmful AI systems to compete with people for survival. Many movies have been produced trying to explain the possible threats of AI systems competing with the mankind. The application of AI in the military is a major concern since it emphasizes on creating systems capable of harming mankind. Militaries across the world apply intelligent drones for security purposes. There are also security robots that are programmed to detect threats and are also empowered to take actions to neutralize specific threats. However, the application of these robots has received opposition, and this has delayed their application (Del Monte 2012). In case they are applied, their continuous development might lead to the creation of harmful systems competing with mankind.
Singularity is a concept coined by Vernor Vinge that suggests that AI will overtake human intelligence. According to this concept, AI invention will grow technology giving learning systems enormous powers over human beings. When this situation is reached, human beings will lose control of the machines making the former a slave of the latter. The phenomenon is feared as it would be a threat to humanity. Currently, machines are not self-directing as they depend on instructions from human beings. Singularity theorists suggest that machines will develop consciousness making them self-directing and this would cause rivalry between them and human beings. Conscious machines would lead to self-improvement systems that can build enormous capabilities beyond the comprehension of human beings. These alterations would cause far-reaching alterations in human civilization.
Based on the literature review AI remains a contentious issue due to its aspects that are hard to understand and predict. The two types of AI have negative and positive impacts, and they continue to experience major changes. People interact with AI in different aspects of lives, and this makes the technology an important part of our lives. The weak AI is more relevant to the people than the strong AI. This relevance stems from the fact that the strong AI is an anticipated technology that is still being improved aimed at creating learning machines that are equivalent or better than human intelligence. Despite it being a future technology, the continuous improvement of the weak AI is hinting proximity to the realization of the strong intelligence. The impact of AI on employment is diverse as it has both negative and positive ramifications. The technology can improve the productivity of skilled employees as it supports knowledge synthesis. On the other hand, AI is a threat to unskilled employees as the application of AI leads to process automation that reduces human involvement in the business process.
There are many myths about AI that need to be clarified through research. Most studies do not agree that AI leads to human extinction due to the possibility of the creation of conscious learning machines capable of initiating self-improvements that would make them better than human beings. Despite many studies not agreeing that AI will outdo human intelligence, some studies discuss some AI features that show the possibility of learning systems overpowering the human race. For instance, the development of a security system that can identify a threat and initiate measures to neutralize the threat is a prototype of systems foreseen in technological singularity. These and other innovations are increasing the possibility of a singularity occurring.
Different singularity theorists have come up with different perspectives regarding the future of AI. Some believe that AI will have grave outcomes that will be a threat to mankind. Stephen Hawking’s singularity theory paints a dark future of AI. According to Hawking, the creation of full artificial intelligence would mark the end of mankind. He says that the basic AI that has been developed so far has proved to be useful as it has improved human life. His greatest fear is the creation of more AI capacities that can surpass human intelligence.
Murray Shanahan is another singularity theorists, and he gives a broad view on technological singularity. He argues that human-level AI is hard to achieve but theoretically possible. The realization of this level of AI would have diverse implications and development to superintelligence would be swift. Superintelligence would create an existential opportunity as well as an existential threat to mankind (Shanahan 2015). In other words, the superintelligence would create an opportunity to improve people’s lives and also be a threat to the same people in case AI systems go rogue.
Danko Nikolic presents a different view on the future of AI. He argues that AI can never surpass the human intelligence. He says that human being’s basic learning tools are contained in the genes that have been evolving towards improvement for billions of years. According to him, it is impossible for machines to acquire similar capabilities (Callaghan, Miller, Yampolskiy & Armstrong 2017). Nikolic, therefore, paints a positive picture of AI where mankind is in full control of learning machines.
It is clear the views of various singularity theorists regarding the future of AI vary. The danger of AI having negative outcomes in future is real. Chances of creating super intelligence from the self-improvement of human-level AI are high. Supper intelligence would have both negative and positive outcomes. The creation of superintelligence is pegged on the development of AI that is equivalent to human intelligence. It is however uncertain that human-level AI will be created since the human thinking process is unique and it is next to impossible for scientists to create a similar learning machine. Chances are therefore high that singularity may never occur.
Artificial intelligence and learning systems have a great impact in today’s world. Computer scientists are striving to create AI systems that are closer to human intelligence. There are two types of AI namely the weak AI and the strong AI. The weak AI is a learning machine created to perform a task or a small number of tasks. On the other hand, a strong AI is a learning system equivalent to human intelligence. This AI is still a work in progress and scientists are hoping to introduce it years to come. The future of AI is uncertain, and different scientists have made diverse predictions about this technology. Technological singularity is a concept that seeks to explain the future of AI. Some singularity theorists are of the view that AI will surpass human intelligence while others think that human intelligence is too strong and unique to be surpassed by AI.
Bench-Capon, T.J. & Dunne, P.E., 2007. ‘Argumentation in artificial intelligence.’ Artificial intelligence, vol.171,no.15, pp.619-641.
Brooks, R.A., 1991. ‘Intelligence without representation.’ Artificial intelligence, vol.47, no.3, pp.139-159.
Callaghan, V., Miller, J., Yampolskiy, R., & Armstrong, S. 2017. The Technological Singularity: Managing the Journey. Berlin, Heidelberg Springer.
Chrisley, R., 2003. ‘Embodied artificial intelligence.’ Artificial Intelligence 149, 131–150.
Del Monte, L. A. 2012. Unraveling the universe’s mysteries: explore sciences’ most baffling mysteries, including the Big Bang’s origin, time travel, dark energy, humankind’s fate, and more. Louis A. Del Monte.
Jacquette, D., 1993. ‘Pollock on token physicalism, agent materialism and strong artificial intelligence.’ International Studies in the Philosophy of Science 7, 127–140.
Lawless, W.F., Mittu, R., Russell, S., & Sofge, D., 2017. ‘Autonomy and artificial intelligence: A Threat or Savior?’ Springer International Publishing, vol.4, no.2, pp. 1–316
Shanahan, M. 2015. The technological singularity. Cambridge, Massachusetts : The MIT Press,