Definition of Artificial Intelligence

Seyedmahmoudian et al. (2016) have defined AI (Artificial Intelligence) as a constellation of various technologies including processing of natural language, reasoning and machine learning. On the other hand, Hovy and Spruit (2016) argued by characterizing AI as a concept of engineering and science that enables intelligence in programs and commuters. AI can be therefore demonstrated as a field of technology that includes intelligent thinking in robots, computers or program providing the computers with the human like thinking capability. The concept of AI has been developed based on the nature of thinking, processing, analyzing and decision making skills followed by human brains.

According to Russell, Dewey and Tegmark (2015) Broussard (2015), several decades of continuous study and research on AI have increased the application of AI in daily life and in various inventions. Wenger (2014) pointed out that AI significantly impacted the human life, daily commodities and the society. The discovery of AI is considered as greatest achievement of mankind. Ruths and Pfeffer (2014) claimed that the developing the ability of machines to mimic the human intelligence, surpassing the thoughts can be easily related to science fiction. Although Müller and Bostrom (2016) warned that AI could be last achievement of human race until, necessary action and procedure can be developed for avoiding the risks.

Machine learning is the common technology that has been widely used and practices in mostly all the inventions and researches related to AI. Ghahramani (2015) showed that transportation is one of the first domain that that aims at integrating AI for improving the safety and reliability of the people on road. The development of smart car and driverless car have paved the way for intelligence cars that can detect the traffic on the road and driving the car based on their decision making capabilities.

Russell, Dewey and Tegmark (2015) claimed that the application of the AI in research work can eliminate the possibility of errors while ensuring the accuracy and reliability of information after analysis. Müller and Bostrom (2016) pointed out AI is being applied by NASA astronauts for better understanding and exploration of space. In daily application, the use of AI has been provided with the potential of easing the human with the capability of learning, reasoning and perception. Seyedmahmoudian et al. (2016) provided example by stating the Cortana and Siri embedded on our Smartphone helps the human in guiding with the daily activities. Russell, Dewey and Tegmark (2015) illustrated that when someone takes pictures, technology of AI has been used for identifying the particular person on the picture for tagging.

Further, the development of digital assistance had allowed in saving financial amount on human resources. In addition to that, the application of AI in organization would provide the same output as human while reducing the chances of errors, with high speed and eliminating biasness due to no sense of emotion and rational thinking. Hovy and Spruit (2016) illustrated that various monotonous, repetitive and lengthy tasks are being carried out by humans that creates errors and takes huge time. Therefore, application of AI for carrying out the monotonous tasks that allowed in multitasking and time and speed can be set as parameter unlike the human.

Social Impact of AI

Furthermore, AI has provided the human race with significant opportunities for enhancing the life and process that are being naturally followed. Russell, Dewey and Tegmark (2015) showed that unlike the human, AI machines require no refreshments or breaks and cannot be distracted from work. Russell, Dewey and Tegmark (2015) claimed that AI is being widely used in medical science for detecting and monitoring the neurological behavior and simulate the function of the human brain in medical practices. Therefore, it has been identified that the application of AI in various operations and processes, and industries helps in easing the work of human.

In spite of the AI have overcome the limitations of human capability in various domains , AI fails to providing several critical judgment and specifications as humans. Russell, Dewey and Tegmark (2015) showed that the application of AI in various domains of life has allowed the machines to follow the goals generated by humans in unconstraint way. The cutting edge technology fails to decide between wrong and right sets of behavior. The wide expansion and rise of AI in various domains have eased human activities but raised various issues, problems ad ethical considerations. Seyedmahmoudian et al. (2016) claimed that the AI machines are deprived of the moral status and considerations. In various field of engineering including medical or manufacturing, accepting and conducting the moral set of rules and acts are essential. Broussard (2015) claimed than disputes found regarding the abortions does not agrees with the moral and ethical status of the embryo. AI applied on the medical field fails to distinguish between the ethical and practical consideration of practice. Therefore, the significant question raises to what extent AI needs to be integrated and used in various aspect and processes of work.

 

Figure 1: Rich Picture on Social Impact of AI

(Source: Developed by Author in MS Visio)

In the above rich picture various scenarios of work and society has been highlighted that ethical and moral issues are associated with various aspect of artificial intelligence. The picture illustrated that with the aim of including AI in the manufacturing process and industries for automating the operations and work have increased the complexity of the physical work at the industries. Furthermore, initiating driverless cars means unemployment of thousands of jobs. But, considering the ethical aspect of reducing accidents, AI cars provides justified options. Wenger (2014) pointed out that the current economic systems runs depending in the compensation raised that are generated based on the hourly wages of the employees. Ruths and Pfeffer (2014) claimed that application of AI significantly cuts down the dependency of the organization on the manpower that reduces distribution of the revenue to a smaller number of people. Müller and Bostrom (2016) questioned in the era of AI machines, what ethical considerations should be taken for distributing the wealth generated by the machines.

Machine learning is the approach applied in the machines for developing artificial intelligence in their actions. Ghahramani (2015) showed that in the training phase, the AI machines are provided with required environment, desired output and input for detecting the results. Russell, Dewey and Tegmark (2015) claimed that unlike the human the AI machines can be followed. Müller and Bostrom (2016) have provided example by stating that providing random dotes in from of the machines world detect any objects that are not here thus compromising the efficiency and security of performance.

Problem Identification

According to Seyedmahmoudian et al. (2016) the application of AI in machines helps in improving the processing capacity and speed but the results cannot be often trusted and neutral. Hovy and Spruit (2016) showed that in an experiment carried out by Google with the photo services, for predicting potential criminals, the AI machines have showed black people and provided biased results. The fact that AI machines, learning environment and training rules are developed by humans who are judgmental and biased during the work often in biased results and output in the AI machines.

Wenger (2014) claimed that Artificial Intelligence has created both excitement and fear from the inception of machine creation. There exist various controversies and debates over the legalities and ethics that need to be followed while working and employing AI. Ruths and Pfeffer (2014) claimed that the in spite of the superfast and flawless generation of coding, AI could provide wrong results in many instances. Müller and Bostrom (2016) showed that in a experiment conducted on 2015, AI have recognize a alternative black and yellow lines as a school bus. Researcher has claimed that in the commercial development product, AI can generate significant amount of negative unexpected outcomes.  

Ghahramani (2015) illustrated that current study has been undertaken with AI research and study while developing the AI with ethical code and analyzing the behavior in a much simpler setting with the humans intelligence are smarter than the AI machines. Ghahramani (2015) claimed that in a recent study have developed an AI for analyzing and simulated the scenario of fertilizer on the population and global temperature. In the experiment, AI showed low dangers and errors, when the AI machines was introduces to only three chemicals.

According to Estevez (2017) the risks and ethical considerations cannot be fully eliminated from the AI machines. But, the impact of the risks can be limited and can be developed into manageable conditions while safeguarding various regularities. Müller and Bostrom (2016)  showed that adequate safeguarding against ethical considerations can be developed while defining a particular AI algorithm and testing protocol and enhancing the validation standard during the testing procedure. Furthermore the initiative for mitigating the ethical issues and risks associated with AI differs with various industries, environment and particular AI machines and operations. Wenger (2014) showed that it is essential to understand that human judgment and involvement cannot be completely eliminated from any process or aspect. There is no denying the fact that AI does assist the humans to easy the work but, human integration and involvement is essential to eliminate the ethical aspect from AI.

Reference 

Broussard, M., 2015. Artificial intelligence for investigative reporting: Using an expert system to enhance journalists’ ability to discover original public affairs stories. Digital Journalism, 3(6), pp.814-831.

Estevez, P.A., 2017. Ethics and the Social Impact of Computational Intelligence [President’s Message]. IEEE Computational Intelligence Magazine, 12(3), pp.4-5.

Ghahramani, Z., 2015. Probabilistic machine learning and artificial intelligence. Nature, 521(7553), p.452.

Hovy, D. and Spruit, S.L., 2016, August. The social impact of natural language processing. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Vol. 2, pp. 591-598).

Müller, V.C. and Bostrom, N., 2016. Future progress in artificial intelligence: A survey of expert opinion. In Fundamental issues of artificial intelligence (pp. 553-570). Springer International Publishing.

Müller, V.C. and Bostrom, N., 2016. Future progress in artificial intelligence: A survey of expert opinion. In Fundamental issues of artificial intelligence (pp. 553-570). Springer International Publishing.

Russell, S., Dewey, D. and Tegmark, M., 2015. Research priorities for robust and beneficial artificial intelligence. Ai Magazine, 36(4), pp.105-114.

Russell, S., Dewey, D. and Tegmark, M., 2015. Research priorities for robust and beneficial artificial intelligence. Ai Magazine, 36(4), pp.105-114.

Ruths, D. and Pfeffer, J., 2014. Social media for large studies of behavior. Science, 346(6213), pp.1063-1064.

Seyedmahmoudian, M., Horan, B., Soon, T.K., Rahmani, R., Oo, A.M.T., Mekhilef, S. and Stojcevski, A., 2016. State of the art artificial intelligence-based MPPT techniques for mitigating partial shading effects on PV systems–A review. Renewable and Sustainable Energy Reviews, 64, pp.435-455.

Wenger, E., 2014. Artificial intelligence and tutoring systems: computational and cognitive approaches to the communication of knowledge. Morgan Kaufmann.

Wenger, E., 2014. Artificial intelligence and tutoring systems: computational and cognitive approaches to the communication of knowledge. Morgan Kaufmann.