Artificial+Intelligence

toc =**Introduction **=

According to Encyclopedia Britannica Artificial Intelligence can be defined as, “the ability of a digital [|computer]  or computer-controlled [|robot]  to perform tasks commonly associated with intelligent beings” [1]. Artificial intelligence is an emerging field of technology. Companies are racing around the world to become the first to create an intelligent artificial being because they could be wildly efficient for economical, and technological aspects of business. Having a super computer that can test and improve on itself would be a huge advantage to a company or a nation. Large companies have started working on the project including: Apple, IBM, Google, and Microsoft. The creation of Artificial Intelligence is a large stepping stone for computing power. Future developments will include classifications of artificial intelligence, current models of artificial intelligence, and philosophical perspectives on the creation of an artificial being.

=**Hard Vs. Soft AI **=

It is important to make a distinction regarding the types of artificial intelligence. Soft or weak artificial intelligence is a computing system that is narrow and focuses on one task. While hard artificial intelligence is a machine with consciousness, sentience, and a mind. Hard artificial intelligence would be able to solve a large array of problems, and can be defined as a machine that is programmed to become self-conscious, and have the ability to add/subtract code onto itself [2]. All current artificial intelligence systems are considered soft artificial intelligence at best. An example of a soft artificial intelligence is Alexa or Siri. These systems are pre-programmed and don’t have any kind of self-sentience that would be required for a hard artificial intelligence. Researchers are much more familiar with soft AI because it has already been developed. Ray Kurzweil, an expert in artificial intelligence, believes that technological capabilities of machines will outmatch human abilities by 2029 [3]. This is the development of hard artificial intelligence.

=**Soft AI **=

There are a series of tests that engineers have placed in front their machines to test their capabilities. The Jeopardy test challenges a machine to beat a human champion in a game of jeopardy [4]. Jeopardy is a game that tests knowledge and recollection of information. For a computer to be able to win the game it would have to rival a human in language with speed accuracy and confidence. The game challenged Watson’s ability to pick up on irony, riddles, and other human language complexities [4]. Watson passed the Jeopardy Test in 2011 beating two human champions. Another example of a soft artificial intelligence system, which was created by Ohlsson and team, was an artificial intelligence. His team tested it with verbal IQ tests that are aimed towards 4 year olds. Their artificial intelligence system scored higher than the average 4-year-old [5].

The use of artificial intelligence is diverse. Researchers have used AI to read human genome sequences with quality results [6]. What used to take years to study a single biomolecule or protein, is now sped up exponentially. A single experiment using the new artificial intelligence system can create up to 3 gigabytes of data. This is the first step in the road towards more efficient medicine, as it is still very difficult to analyze this data. Other uses of AI so far include: the use of AI to sift through law journals for lawyers, the creation of games, generating images, aid in finances, and the control of traffic more efficiently [7][8][9][10][11]. =**Hard AI **=

There is a race around the world to be the first to create a hard AI. To be able to have a hard-artificial intelligence it would require the machine to pass the imitation game also known as the Turing Test. The imitation game is played with three players a man a woman and an interrogator, the interrogator asks questions to the man and woman without ever seeing them and then determines which one is a man and which is a woman. Turing raises the dilemma of how would a machine fare in the imitation game [12]. This is seen within the engineering community as the challenge to beat. The machine would have to be able to understand language and have the ability to infer from the data provided to decide which individual is which.

There are many problems in creating an artificial person though, and throughout the process of creation it is imperative to properly manage the machine for the safety of humans. A self-aware individual typically has these goals: self-preservation, goal-content integrity, cognitive enhancement, technological perfection, and resource acquisition. Lawrence and teams advises that to properly manage the new technology scientists would need to figure out how to effectively control them before letting them loose into society [13]. One way to ensure management of the artificial beings is to ensure that their goal setting is similar to our own. The real fear of developing an AI person would be that it could have radically different goals then our own, which leaves us in the dark as to what they may attempt to achieve. Another very difficult aspect in producing an artificial person is mimicking human consciousness so the new being would be able to properly relate, and communicate with humans. Humans have many elements that all play into our consciousness, and they would all need to be mimicked for a machine to be able to relate with humans. These elements include: empathy, physical world capability’s (senses), communication, memory, knowledge Corpus (like Wikipedia), explainability, consciousness feedback loops, and dreams for reflection [14]. The trick in creating an AI would be finding the correct mixture of all of these elements. Banerjee and team came up with the modules and how they should work together to create a being that is able to properly communicate with humans. Scientists can expect much trial and error in these experiments. It is imperative to keep this machines in a safe experimental area with proper monitoring by human supervision [14].

<span style="font-family: Arial,sans-serif; font-size: 10pt;">Another major problem in creating a new artificial intelligence is being able to replicate how a human approaches and attacks a problem. Humans form a mental model to consider the problem and this is exceedingly difficult to replicate within machine learning. Humans are able to manipulate problems while issues are only partially understood and then through selective thinking problems can be solved. The method that is currently in use for machine learning and is currently being improved upon is machines search and search until answers can be found such as IBM Watson playing jeopardy. Although machines are able to search extremely quickly, they still need to be able to infer. This is the next major step in creating a hard artificial intelligence. For a machine to be practical they need to use knowledge and insight together to better solve problems. Humans use a heuristic approach to problem solving by focusing on practicality, and machines still lack this [15].

=**<span style="font-family: Arial,sans-serif; font-size: 10pt;">Conclusion **=

<span style="font-family: Arial,sans-serif; font-size: 10pt;">The invention of hard artificial intelligence is still in its infancy. The invention of soft artificial intelligence is a great platform to start with, but there are still many loops to cross to create a self-conscious and sentient being with the capabilities to improve upon itself. Soft artificial intelligence is relatively understood, but still needs to gain many more features and be tested more fully before the technology will have major impacts on our lives. Hard artificial intelligence is not very well understood at all. The creation of hard artificial intelligence has been coined by the philosophical term the singularity [16]. The singularity in philosophical terms means there is no possible outcome that can be seen past this event. The impacts of this event could be massive towards society. Tethering the invention to a philosophical term where nothing is known after the event shows how little the experts know about what this invention could entail. The invention of this machine is what needs to be studied the most going forward.

=**<span style="font-family: Arial,sans-serif; font-size: 10pt;">Bibliography **=

[1] B.J. Copeland (2018), "Artificial Intelligence," //Encyclopedia Britannica.//

<span style="font-family: Arial,sans-serif; font-size: 13.3333px;">[2] Anonymous, "Artificial Intelligence; Investigators from Research Institute Zero in on Artificial Intelligence (Artificial intelligence in service-oriented software design //<span style="font-family: Arial,sans-serif; font-size: 13.3333px;">)," Computers, Networks & Communications //<span style="font-family: Arial,sans-serif; font-size: 13.3333px;">, pp. 227, 2016.

[3] Ray Kurzweil (2017), "Human Level AI is Right Around the Corner- or Hundreds of Years Away," //IEEE Spectrum.//

<span style="font-family: Arial,sans-serif; font-size: 10pt;">[4] R. Chandrasekar, (2014, March). "Elementary? Question Answering, IBM’s Watson, and the Jeopardy! Challenge, //<span style="font-family: Arial,sans-serif; font-size: 10pt;">" Resonance. //

<span style="font-family: Arial,sans-serif; font-size: 10pt;">[5] E. Brown (2012), "Watson: The jeopardy! challenge and beyond,". DOI: 10.1145/2348283.2348446.

<span style="font-family: Arial,sans-serif; font-size: 10pt;">[6] S. Ohlsson et al, "Measuring an artificial intelligence system's performance on a Verbal IQ test for young children," Journal of Experimental & Theoretical Artificial Intelligence, vol. 29, (4), pp. 679, 2017.

<span style="color: #333333; font-family: Arial,sans-serif; font-size: 10pt;">[7] Lawton, George (2012, July). "Researcher uses AI to analyze human genomes". //IEEE intelligent systems// (1541-1672), 27 (4), p. 6.

<span style="color: #333333; font-family: Arial,sans-serif; font-size: 10pt;">[8] G. Lawton (2011, September). "Software uses AI in place of lawyers". //IEEE intelligent systems// (1541-1672), 26 (5), p. 8.

<span style="color: #333333; font-family: Arial,sans-serif; font-size: 10pt;">[9] G.Lawton (2012, May). "Researchers use AI to build games". //IEEE intelligent systems// (1541-1672), 27 (3), p. 6.

<span style="color: #333333; font-family: Arial,sans-serif; font-size: 10pt;">[10] D. Castelvecchi, “Atronomers explore uses for AI-generated images,” Nature, vol. 542,(7639), pp 16-17,2017.

<span style="color: #333333; font-family: Arial,sans-serif; font-size: 10pt;">[11] Q. Fan //et al//, "Use AI Technology to Analyse Corporate Goods Price Index," //Journal of Software,// vol. 6, //(11),// 2011.[10] G. Lawton, "Engineers use AI to make traffic control more effective," //IEEE Intelligent Systems,// vol. 27, //(6),// pp. 5, 2012.

<span style="font-family: Arial,sans-serif; font-size: 10pt;">[12] A.M. Turing “Computing Machinery and Intelligence” //Mind// vol. 49, pp 433-460, 1950.

<span style="font-family: Arial,sans-serif; font-size: 10pt;">[13] LAWRENCE, et. Al. (2016). Artificial intelligence. Cambridge Quarterly of Healthcare Ethics, 25(2), 250-261.

<span style="font-family: Arial,sans-serif; font-size: 10pt;">[14] S. Banerjee, "A framework for designing compassionate and ethical artificial intelligence and artificial consciousness," //PeerJ PrePrints//, 2018.

<span style="font-family: Arial,sans-serif; font-size: 10pt;">[15] S. Singh and S. Singh, “Artificial Intelligence,” //International Journal Computer Applications//, vol. 6, (6), 2010.

<span style="font-family: Arial,sans-serif; font-size: 10pt;">[16] Anonymous "Artificial Intelligence; Investigators from Research Institute Zero in on Artificial Intelligence (Artificial intelligence in service-oriented software design//)," Computers, Networks & Communications//, pp. 227, 2016.

<span style="font-family: Arial,sans-serif; font-size: 10pt;">[17] T. Manzini, S. Ellis and J. Hendler, "A Play on Words: Using Cognitive Computing as a Basis for AI Solvers in Word Puzzles," //Journal of Artificial General Intelligence,// vol. 6, (1), pp. 111-129, 2015.