As discussed in my previous article, Artificial Intelligence (AI) promises to change our lives in all industry sectors. However, it is still an ongoing debate whether AI will be a boon or bane to our society. Many technology experts have already expressed their concerns regarding possible impacts of AI on our society.[1] In this article, I would like to discuss the societal challenges and limitations that are associated with AI.

1. High costs of AI machines

One of the main challenges of AI is the cost of smart devices and machines. Due to the complexity in programming and thus the necessity of experienced programmers to build these devices and machines make the development very expensive. Hence, small business sectors cannot even afford the basic AI devices due to their low capital investments. However, for larger companies, smarter AI devices will be needed depending on the type of project and the corresponding AI devices will be very costly. In addition to the initial construction cost, the maintenance of AI machines, as well as updating of programs will also require huge costs.[2]

2. Requirement of huge high-quality data

AI machines need greater volumes and sources of quality data in order to enable the capabilities of AI for providing meaningful tasks and efficient results. Even though huge amounts of data can be collected, the integration of these data and the optimization of algorithms remain incredibly hard. The significant limitations are visible in sectors such as healthcare due to the inadequate amount of data, lack of sufficient quality information and difficulty in aggregating data across multiple medical record systems. For example, a medical practitioner cannot predict the outcome of a rare disease and the disease progression with the help of AI devices due to the lack of sufficient patient medical records regarding the disease and its symptoms.[3][4][5]

3. Declining cognitive development

Human beings have the capability to learn and improve their skills from each and every experience, and hence they learn how to respond to unique situations. AI devices make humans depend more on technology, which reduces our endless self-development possibilities, lateral thinking, and multitasking abilities. The overdependence on machines will slowly reduce our thinking abilities, mental capacities and thus affect our cognitive development. Hence, the AI machine replacement will affect our overall development and mental health.[6][7]

4. Lack of human intelligence and creativity

AI can never emulate the human mind in terms of creative thinking, common sense and emotional intelligence. AI machines cannot replicate the intuitive abilities such as judgment or decision-making skills based on previous experiences and knowledge, and the inherent human abilities. However, AI devices can perform only in such a way they are programmed to work, which has limitations to perform better with experience and to change their responses differently even if the situation demands to act differently. Since the creative thinking and intelligence of AI devices will be restricted only to the programmed algorithms, there is no possibility to train AI devices the concept of thinking “out of the box” for improving its intelligence to emulate human intelligence and creativity. [8][9][10][11]

5. Privacy and security

Privacy and security are important aspects that need to be addressed. The AI devices can track and analyze our each and every online behavior as well as offline activities with the help of AI cameras through the facial recognition algorithm when we do our daily business. This AI behavioral analysis and its decision-making power may lead to invasion of privacy and will eventually turn into social oppression. Another greatest threat with AI is the decision making in certain circumstances due to its lack of human intelligence and its limitations to act only with the provided data. This may lead to the possibility of massive destruction when the control of AI devices is given in wrong hands to program them.[12][13][14]

6. Unemployment

Although AI dramatically improves our quality of life, it will cause many social and economic consequences due to its impact on employment and workforce in the coming years. The job automation due to AI will affect different sectors such as transportation, finance, insurance, electronic commerce, legal professions, recruiting and will also displace many other low-skilled jobs. In addition, the replacement of robots for jobs in the manufacturing sector, the concept of driverless cars, AI-powered tutors, and AI-powered consultants clearly convey us the most dangerous impact of AI in the future in the form of unemployment. Though many jobs are created through the advancement of AI technology, there will also be job loss or jobs at risk due to the AI machine replacement for the above-mentioned business sectors which need less skills for repetitive tasks, and this could extend to new levels. [15][16][17]

7. Ethical and legal issues

Another important concern regarding AI is the ethical and legal questions related to the creation of replicas of human beings or intelligence. Ethical and legal concerns associated with AI, hinder the acceptance and trust of AI machines in many sectors have been raised. This includes major issues such as algorithmic fairness, privacy, transparency, the possibility of discrimination and bias, accountability of AI decisions, explainability of the algorithm, interpretability of AI results, wealth management, and economic impact due to job displacement and job loss. In addition, human rights due to unemployment, safety and security of humans by the use of autonomous weapons and autonomous vehicles, cybersecurity, and human rights to know about AI results and algorithmic decisions are the further concerns about AI.[18][19][20][21]

8. Change of organizational culture

Each organization and business sector have its own unique company culture, which includes the collection of values, beliefs, principles, norms, behaviors, assumptions, habits, environment and location of the organizational members. The organizational culture influences the members of each organization in people interaction, togetherness, trust, cultural awareness, team cohesiveness, and goal achievement and thereby improve their total performance and productivity. As a human being, interpersonal interactions, mutual trust, and experience at work are considered as essential to our well-being and professional satisfaction. Hence maintaining and improving a human workforce and AI workforce on the same platform is too tricky to employ.[22][23]

9. Machines supersede human

The possibility of machines overruling humans is another anxiety and fear regarding AI. Until now, humans are the masters of AI machines and the main concern about AI is the possibility of smarter AI machines dominating and ruling human and thus the whole world. This may lead to a power shift from human to AI due to AI dominance, and there may be a chance of AI control in our society. Hence, the emerging AI technology may lead to the deterioration of humanity. In addition, the treatment of human only as a commodity may lead to unpredictable difficulties in the future.[24]

Due to the above-mentioned aspects, each business sector needs to carefully study the benefits and risks of AI before employing AI machines. Let us look forward toward a future, in which AI machines will lead and support humans to live a better life by addressing the societal risks and challenges involved.

References

[1] Stephen Hawking et al., The Independent, 2014, 2014, 9313474.

[2] https://en.wikipedia.org/wiki/Artificial_intelligence

[3] Thomas M Maddox et al., JAMA, 321, 2019, 31-32.

[4] Peter M A van Ooijen, Artificial Intelligence in Medical Imaging, 2019, 247-255.

[5] Jyoti Soni et al., International Journal of Computer Applications, 17, 2011, 43-48.

[6] Helbing D. et al., Towards Digital Enlightenment, 2019, 73-98.

[7] James Hendler et al., Social Machines: The coming Collision of Artificial Intelligence, Social Networking, and Humanity, Apress. 2016.

[8] Mitchell M Waldrop, PNAS, 116, 2019, 1074-1077.

[9] Jean-Charles Pomerol, European Journal of Operational Research, 99, 1997, 3-25

[10] Margaret A Boden, Artificial Intelligence,103, 1998, 347-356.

[11] Sonoo Thadaney Israni et al., JAMA, 321, 2019, 29-30.

[12] Vicenc Torra et al., Data Science in Practice: Studies in Big data, 46, 2019, 121-132.

[13] Akbar Khanan et al., Smart Technologies and Innovation for a Sustainable Future, 2019, 55-61.

[14] Mark Ryan, The ORBIT Journal, 2, 2019, 1-33

[15] Dirk Helbing, Towards Digital Enlightenment, 2019, 47-72.

[16] Fabio Massimo Zanzotto, Journal of Artificial Intelligence Research, 64, 2019, 243-252.

[17] Toby Walsh, International Journal of Automation and Computing, 15, 2018, 637-642

[18] Corinne Cath, Philosophical Transactions A Mathematical, Physical Engineering Sciences, 376, 2018, 1-8.

[19] Greg Adamson et al., Proceedings of the IEEE, 107, 2019, 518-525.

[20] Sean Kanuck, Ethics & International Affairs, 33, 2019, 3-12.

[21] Alan Winfield, Nature Electronics, 2, 2019, 46-48

[22] David Brougham et al., Journal of Management & Organization, 24, 2018, 239-257.

[23] Kun Yu et al., IUI 2019 Proceedings, 2019, 460-468.

[24] Andreas Kaplan et al., Business Horizons, 62, 2019, 15-25.