- 431 Downloads
Artificial Intelligence (often referred to as AI) is a field in computer science that is concerned with the automation of intelligence and the enablement of machines to achieve complex tasks in complex environments. This definition is an augmentation of two preexisting commonplace AI definitions (Goebel et al. 2016; Luger 2005).
AI is an umbrella that has many subdisciplines, big data analytics is one of them. The traditional promise of machine intelligence is being partially rekindled into a new business intelligence promise through big data analytics.
This entry covers AI, and its multiple subdisciplines.
AI is a field that is built on centuries of thought; however, it became a recognized field for only over 70 years or so. AI is challenged in many ways, identifying what’s artificial versus what is real can be tricky in some cases, for example: “A tsunami is a large wave in an ocean caused by an earthquake or a landslide. Natural tsunamis occur from time to time. You could imagine an artificial tsunami that was made by humans, by exploding a bomb in the ocean for instance, yet, it still qualifies as a tsunami. One could also imagine fake tsunamis: “using computer graphics, or natural, for example, a mirage that looks like a tsunami but is not one.” (Poole and Mackworth 2010).
However, intelligence is arguably different: you cannot create an illusion of intelligence or fake it. When a machine acts intelligently, it is then intelligent. There is no known way that a machine would demonstrate intelligence randomly.
The field of AI continuously poses a series of questions: How to define or observe intelligence? Is AI safe? Can machines achieve superintelligence? among many other questions. In his famous manuscript, “Computing Machinery and Intelligence” (Turing 1950), Turing paved the way for many scientists to think about AI through answering the following: Can Machines think? To be able to imitate, replicate, or augment human intelligence, it is crucial to first understand what intelligence exactly means. For that, AI becomes a field that overlaps other areas of study, such as biology (the ability to understand the human brain and nervous system); philosophy is another field that has been highly concerned with AI (understanding how AI would affect the future of humanity – among many other philosophical discussions).
Machine Learning: is when intelligent agent learn by exploring their surroundings and while figuring out what actions are the most rewarding.
Neural Networks: are a learning paradigm inspired by the human nervous system. In neural networks, information is processed by a set of interconnected nodes called neurons.
Genetic Algorithms (GA): is a method that finds a solution or an approximation to the solution for optimization and search problems. GAs use biological techniques such as mutation, crossover, and inheritance.
Natural Language Processing (NLP): is a discipline that deals with linguistic interactions between humans and computers. It is an approach dedicated to improving the human-computer interaction. This approach is usually used for audio recognition.
Knowledge-based systems (KBS): are intelligent systems that reflect the knowledge of a proficient person, also referred to as expert systems. KBS are known to be one of the earliest disciplines of modern AI.
Computer Vision: is a discipline that is concerned with injecting intelligence to enforce the ability of perceiving objects. It occurs when the computer captures and analyzes images of the 3D world. This includes making the computer recognize objects in real-time.
Robotics: is a central field of AI that deals with building machines that imitate human actions and reactions. Robots in some cases have human features such as arms and legs, and in many other cases, are far from how humans look like. Robots are referred to as intelligent agents in some instances.
Data Science and Advanced Analytics: is a discipline that aims to increase the level of data-driven decision-making and providing improved descriptive and predictive pointers. This entry has been the focus of recent business AI applications, to the degree that many interchangeably (wrongly though) refer to it as AI. Many organizations are adopting this area of research and development. It has been used in many domains (such as healthcare, government, and banking). Intelligence methods are applied to structured data, and results are usually presented in what is referred to as a data visualization (using tools such as Tableau, R, SPSS, and PowerBI).
Computer agents are a type of intelligent system that can interact with humans in a realistic manner. They have been known to beat the world’s best chess player and locate hostages in a military operation. A computer agent is an autonomous or semiautonomous entity that can emulate a human. It can be either physical such as a robot or virtual such as an avatar. The ability to learn should be part of any system that claims intelligence.
AI Challenges and Successes
Intelligent agents must be able to adapt to changes in their environment. Such agents, however, have been challenged by many critics and thinkers for many reasons. Major technical and philosophical challenges to AI include: (1) The devaluation of humans: many argue that AI would replace humans in many areas (such as jobs and day-to-day services). (2) The lack of hardware that can support AI’s extensive computations. Although Moore’s law sounds intriguing (which states that the number of registers on an integrated circuit is doubling every year), that is still a fairly slow pace for what AI is expected to require in terms of hardware. (3) The effect of AI: Whenever any improvement in AI is accomplished, it is disregarded as a calculation in a computer that is driven by a set of instructions, and not real intelligence. This was one of the reasons the AI winter occurred (lack of research funding in the field). The field kept providing exploratory studies but there was a lack of real applications to justify the funding. Recently however, with technologies such as Deep Blue and Watson, AI is gaining attention and attracting funding (in academia, government, and the industry). (4) Answering the basic questions of when and how to achieve AI. Some researchers are looking to achieve Artificial General Intelligence (AGI) or Superintelligence, which is a form of intelligence that can continuously learn and replicate human’s thought, understand context, develop emotions, intuitions, fears, hopes, and reasoning skills. That is a much wider goal of AI than the existing Narrow-intelligence, which presents machines that have the ability perform a predefined set of tasks intelligently. Narrow intelligence is currently being deployed in many applications such as driverless cars and intelligent personal assistants. (5) Turing’s list of potential AI roadblocks: presented in his famous paper, those challenges are still deemed relevant (among many other potential challenges).
In spite of the listed major five challenges, AI already presented multiple advantages such as: (1) greater calculation precision, accuracy, and the lack of errors, (2) performing tasks that humans are not able to or ones that are deemed too dangerous (such as space missions, and military operations), (3) accomplishing very complex tasks such as fraud detection, events prediction, and forecasting. Furthermore, AI had many successful deployments such as: Deep Blue (a chess computer), Autonomous Cars (produced by Tesla, Google and other technology, and automotive companies), IBM’s Watson (a jeopardy computer), and Intelligent Personal Assistants (such as Apple’s Siri and Amazon’s Alexa).
AI is a continuously evolving field; it overlaps with multiple other areas of research such as computer science, psychology, math, biology, philosophy, and linguistics. AI is both feared by many due to the challenges listed in this entry and loved by many as well due to its many technological advantages in critical areas of human interest. AI is often referred to as the next big thing, similar to the industrial revolution and the digital age. Regardless of its pros, cons, downfalls, or potential greatness, it is an interesting field that is worth exploring and expanding.
- Goebel, R., Tanaka, Y., & Wolfgang, W. (2016). Lecture notes in artificial intelligence series. In: Proceedings of the ninth conference on artificial general intelligence, New York.Google Scholar
- Luger, G. (2005). Artificial intelligence, structures and strategies for complex problem solving (5th ed.). Addison Wesley, ISBN: 0-321-26318-9.Google Scholar
- Poole, D., & Mackworth, A. (2010). Atificial intelligence: Foundation of computer agents (1st ed.). Cambridge University Press, ISBN: 978-0-511-72946-1.Google Scholar