Artificial Intelligence (AI) is one of the most hyped technologies currently, at it is widely believed to replace humans in some occupations and transform the way work is done in the remaining jobs. According to Wang (2018), an intelligent system is able to work in real-time, open to unexpected tasks, and able to adapt by learning from its experiences. In a recent study (Saukkonen et al., 2019) on the perceptions about emerging technologies by HRM professionals we defined that “AI is a field of computer science dedicated to solving cognitive problems commonly associated with human intelligence, such as learning, problem-solving and pattern recognition.”
Promises and development of AI impact to the world of work
AI has been stated to be a “generic” and “ubiquitous” technology within academic research (e.g. Dwivedi et al., 2019; Bifet and Read, 2018), meaning it is likely to appear “everywhere, all the time” and to touch most areas of human activity. The enabling nature of AI – what it has to offer – has been studied more for specific purposes and occupations (Black and Van Esch, 2020; Cope et al., 2020).
AI has the potential to replace human work and judgment in organizations, but it can also augment human capability. Sadiku et al. (2020) highlight the differences between Human ns. non-Human intelligence, and state that a realistic target of AI development would be a system that can integrate the virtues of both humans and machines. Frey and Osborne (2017) claim that “even with recent technological developments, allowing for more sophisticated pattern recognition, human labor will still have a comparative advantage in tasks requiring more complex perception and manipulation”. Not surprisingly, then, a strong stream in the research focuses on technology as a tool for human augmentation (e.g. Valeriani et al., 2019) and human-machine interaction (e.g. Bolton et al., 2018) in task fulfillment and the impact on occupations and processes within them.
AI – a generic or a sector-specific issue – Our recent studies
In a recent study – the outcomes of which are discussed more in detail in the proceedings of ECLAIR (European Conference on the Impact of Artificial Intelligence and Robotics) 2021, we subjected three separate studies (BBA Theses) on AI impacts to a meta-analysis and integrative findings (Saukkonen et. al., 2021).
The three studies touched occupations largely seen to include a strong human understanding and judgment nature. The studies were on AI´s role in the recruitment function of Human Resource Management (Lahti, 2020), Prospective AI usage in Primary Health Care (Anton, 2020), and Scenarios for AI´s role in artistic creative work, more specifically in music composing (Karpov, 2020).
All three occupations were assessed to be “relatively safe” occupations vs. to be taken over by machines (by Frey and Osborne, 2017). HR Managers were calculated to be at a risk of 0.55 % probability to be replaced by machines, medical doctors at a risk of 0.4 to 0.5 % depending on the field of specialization, and, finally, music composers at a 1.5 % risk in this respect. However, all substudies we superimposed over each other highlighted areas where AI can transform and speed up processes within the diverse set of occupations. Thus, if AI will not replace humans, at least it can redefine the contents and the way the work is done. Similarly, in all these studies also potential obstacles to AI adoption were identified.
One of the generic promises of AI is to be able to relentlessly screen a vast data pool looking for patterns and propose choices. This, however, has different meanings depending on the application area. In the medical sphere, AI’s can assist doctors by comparing the task at hand to a vast pool of past cases screened. far more than even the most experienced expert can memorize or have quick access to. However, this AI benefit repeats in the recruitment functions but, in the artistic field (music composing) the solutions need to deviate from prior solutions and AI is used to expand the number of potential solutions rather than narrow them. Likewise, both the artistic field and HRM are using data from limited and format-wise similar data sources, whereas the medical field combines data from sources different both in sources and format.
In all cases, the AI´s role is (currently) preparing the final decision to be taken by a human, largely due to the uniqueness of cases that no algorithm can fully cope with as well as due to legal/ethical considerations of responsibility.
When looking at the challenges or hindrances of AI adoption some generic features and some sector-specific ones emerge. The major sector-crossing obstacle is the unsolved legal and ethical stance to responsibilities of machine-made solutions and/or recommendations. AI output depends on the algorithm written and training data chosen (though the reliability of the data do is expected to improve over as data accumulates), biases and errors are possible.
The chain of reasoning behind the algorithm cannot be currently fully shared between the parties due to the lack of common ground and knowledge on AI by the stakeholders of processes. As a result, naturally, no party (algorithm creators, algorithm users, AI-supported decision-makers) is willing to take full responsibility in cases where AI usage may bring negative outcomes. This is common to the areas of Medical and HRM domains. In the artistic field, the issues of rights and responsibilities concern more the positive outcome, creations made by/with AI, and the intellectual property rights created in the process, partly by machines.
The challenges of AI usage across domains are mostly connected to them (so far) unsolved legal and ethical issues: The borderlines of rights and responsibilities between stakeholders as well as between “a man and the machine” are a major challenge to AI adoption. In the prevailing unclear state of these issues, many potential AI users are likely to opt-out of AI adoption.
Our study indicates that the individuals and organizations deciding on the usage/non-usage of AI should not only understand the generic (positive and negative) characteristics of AI but also the issues specific to their context. These considerations should be worked out in cooperation with different stakeholders of business processes, and this discussion should also involve AI developers.
Our integration of findings of the three sub-studies proposes that the deployable and productive AI systems should use the generic principles of AI (e.g. natural language processing, pattern recognition) but for a full effect they should be are tuned for the purpose and context. This creates a go-to-market obstacle to AI system developers, as it affects the way AI solutions can be productized (vs. “projectized”) and scaled up (e.g. as Software-as-a Service= SaaS offering). On the other hand, AI understanding and capability constraints of issue area specialists to participate in limit the proceeding of AI.
by Juha Saukkonen, D.Sc. (Econ.), senior lecturer (management), JAMK Univ. of Applied Sciences, International Business
Anton, K. (2020). Future Foresight into Application of AI Technologies in Finnish Healthcare: Outlining prospective trends of development in primary care by 2030. A Bachelor´s Thesis. Accessible at: https://www.theseus.fi/bitstream/handle/10024/343859/Thesis_Anton_Kirill.pdf?sequence=2&isAllowed=y. Retrieved 29.11.2021.
Bifet, A., & Read, J. (2018). Ubiquitous artificial intelligence and dynamic data streams. In Proceedings of the 12th ACM International Conference on Distributed and Event-Based Systems, DEBS ’18, ACM, New York, NY, USA, pp. 1-6. http://doi.acm.org/10.1145/3210284.3214345
Black, J. S., & van Esch, P. (2020). AI-enabled recruiting: What is it and how should a manager use it?. Business Horizons, 63(2), 215-226.
Bolton, C., Machová, V., Kovacova, M., & Valaskova, K. (2018). The power of human-machine collaboration: Artificial intelligence, business automation, and the smart economy. Economics, Management, and Financial Markets, 13(4), 51-56.
Cope, B., Kalantzis, M., & Searsmith, D. (2020). Artificial intelligence for education: Knowledge and its assessment in AI-enabled learning ecologies. Educational Philosophy and Theory, 1-17.
Dwivedi, Y. K., Hughes, L., Ismagilova, E., Aarts, G., Coombs, C., Crick, T., et al. (2019). Artificial intelligence (AI): Multidisciplinary perspectives on emerging challenges, opportunities, and agenda for research, practice, and policy. International Journal of Information Management. https://doi.org/10.1016/j.ijinfomgt.2019.08.002.
Frey, C. B., & Osborne, M. A. (2013). The future of employment: How susceptible are jobs to automation? Technological Forecasting and Social Change, 114, 254-280.
Karpov, N. (2020). Artificial Intelligence for Music Composing: Future Scenario Analysis. A Bachelor´s Thesis. Accessible at: https://www.theseus.fi/bitstream/handle/10024/339305/Karpov%20Nikita%20Thesis.pdf?sequence=2. Retrieved 29.11.2021
Lahti, W. (2020). Potential of AI Assisted Recruitment in Technology Businesses. A Bachelor´s Thesis. Accessible at: https://www.theseus.fi/bitstream/handle/10024/343483/Lahti%20William.pdf?sequence=2. Retrieved 29.11.2021
Sadiku, M. N., Ashaolu, T. J., & Musa, S. M. (2020): Essence of Human Intelligence. ResearchGate https://www.researchgate.net/publication/348648151_Essence_of_Human_Intelligence. Retrieved 8.2.2021.
Saukkonen, J., Kreus, P., Obermayer, N., Ruiz, Ó. R., & Haaranen, M. (2019). AI, RPA, ML, and Other Emerging Technologies: Anticipating Adoption in the HRM Field. In Proceedings of ECIAIR 2019 European Conference on the Impact of Artificial Intelligence and Robotics. Academic Conferences and Publishing Ltd. Reading, UK. pp. 287-296
Saukkonen, J., Anton, K., Karpov, N. and Lahti, W. (2021). Human Rights, Employee Rights and Copyrights: Parallels of AI Enablers and Obstacles Across Occupations in Human-Centric Domains. In Proceedings of the 3rd European Conference on the Impact of Artificial Intelligence and Robotics ECIAIR 2021, A Virtual Conference Hosted By Iscte –Instituto Universitário de Lisboa 18 -19 November 2021, 166-173. Academic Conferences International. Reading, UK.
Valeriani, D.; Cinel, C. & Poli, R. (2019). Brain-computer interfaces for human augmentation. Brain Science. 2019, 9, 22.
Wang, P. (2008). What Do You Mean by “AI”? In Wang, P., Goertzel, B., and Franklin, S., eds., Artificial General Intelligence 2008. In Proceedings of the First AGI Conference, Frontiers in Artificial Intelligence and Applications, volume 171. Amsterdam, The Netherlands: IOS Press. 362–373.