7. Supposed Consequences of Al in:

Henry Alexander Wittke

Artificial Intelligence, page 54 - 63

An Approach to Assess the Impact on the Information Economy

1. Edition 2020, ISBN print: 978-3-8288-4459-9, ISBN online: 978-3-8288-7480-0,

Tectum, Baden-Baden
Bibliographic information
This Thesis is developed by the results of the analysis of the Information Economy. Since the impact of strong AI on the Information Economy is so massive, the following chapter will seek to establish further the validi ty of the action claim put forward by looking at already occurring im pacts of AI. These short elaborations seek to foster the rather theoretical argument to Machlup’s framework with current evidence. After these supposed consequences of strong AI have been set out, it should be con centrated on the investigation of another hypothesis: (b)To encounter the highly likely effects of AI and in particular strong AI, governments should seek actions to shape the effects of technological changes on the society and facilitate more equal distribution of the positive effects of technical changes. The question here then is, what should governments and even whole so cieties should actually do, which strategies shall be implemented, which actions taken? The discussion on the two hypotheses formulated above shall take place in the next two chapters. 7. Supposed Consequences ofA l Based on the analysis of the information economy (see Chapter 5), it can be assumed that strong AI has the potential to disrupt the main subdivi sions of the information economy and the occupational structures, which result from today’s information societies and will be affected by massive labor market changes resulting from the use of strong AI. Also, the use of AI will result in profound changes in the required skills set for knowledge workers, as well as changing the ways and means to do their jobs. As mentioned in the last chapter, the following section will review the results of this analysis using recent studies and further literature. The first challenge constitutes the fact that there are no isolated studies and statistics with a singular focus on strong AI. One reason for this may be that recent studies and the majority of the literature on AI, in contrast to this work, did not make differentiation of AI into weak and strong AI. Instead, AI is used as an umbrella term; severely limiting the precision of the analysis. In general, AI-driven automation is often cov ered by the same statistics as digitalization or computerization. Thus, it is not possible to show only strong AI within this chapter. 54 Because of the missing data on strong AI, and its possible impact on the economy, statistics on AI-driven automation and computerization are discussed as a relevant phenomenon in this chapter. Based on this it should be assumed that the impact of the ongoing digitalization on socie ty will increase significantly and speed up with strong AI as an accelera tor for these developments in the future. Based on the further literature, the current state of AI is already becoming more deeply integrated into people’s lives, and with further improvements towards strong AI in different industries, it could become the new infrastructure. It will further accelerate the often quoted ongoing industrial revolution (cf. Rotman 2017; Narula 2018; Wahlmuller- Schiller 2017). Especially when the analyzed potential of strong AI in the information economy comes true, a remarkable revolution will be ushered in. Due to the multifold impact of AI and the ongoing digitaliza tion in the information economy, entire socio-economic systems of in formation societies already entered a phase of accelerating transfor mation. Consequently, markets, businesses, government, social welfare, education and employment models will be impacted severely more and more (cf. Krasadakis 2017). It is already predicted that the speed of technological change from the forthcoming AI revolution will rise sharp ly; opening significant opportunities for growth and profitability, as well as novel challenges and competition (cf. Makridakis 2017: 18). Today, it may be challenging to predict exactly how the labor market will change and which jobs will be most immediately affected by AI-driven automation. As mentioned above, AI is not a single technolo gy, but rather a collection of technologies that are applied to specific tasks. This means that the effects of AI will be felt unevenly across the economy. There will be work tasks more easily automated than others, and even there will be jobs, which will be more affected than others. However, based on the current trajectory of AI technology in general, some specific predictions are possible (cf. U.S. Government 2016: 8). AI and robotics are driving rapid and radical workplace transformation across all industries, for small and large companies (cf. EmTech 2018). In ten years, industries will not look anything like today, and in 20 years, much of today's jobs will no longer exist (cf. Braun 2015). The follow ing illustration shows diverse industries that will be affected first: 55 Short Term, Big Bang Long Term, Big Bang Retail ICT& Media Banking £ Insurance Leisure Professional Food Services Education Manufacturing £ Healthcare Real #2 Utilities 0 Agriculture Government Transportation f • Construction 10 Mining. Oil. Gas, 5 1 Chemicals Short Term , Sm all Bang Long Term , Small Bang Remarks: Based on a Report by Deloitte Digital and Heads!. Figure VI: IndustryDisruptionMap. Similar: Deloitte Digital 2015: 1. The illustration shows the impact o f AI across different industry sectors. In the short- and mid-term, AI will significantly impact specific indus tries, such as retail, banking, insurance or education. However, in the long run, AI may have impacted almost every type of industry that exists today. Nevertheless, based on the research so far, it can be assumed that more digitized areas will be affected by strong AI faster than less digitized ones. And the more complex the requirements within an industry are for certain tasks that require, for example, a high level o f creativity, the later strong AI will exert influence. All in all, within these modern industries the world’s top tech companies are in a race to build the best AI and cap ture that massive market means the technology will get better fast and increase its impact on society at all (cf. Maney 2016). “There is little doubt that AI holds enormous potential as comput ers and robots will probably achieve, or come close to, human in telligence over the next twenty years becoming a serious competi tor to all the jobs currently performed by humans and for the first time raising doubt over the end of human supremacy.“ (Makridakis 2017: 2) 56 Based on the literature on Al and computerization, as well as its poten tial impact on the future work dynamics, many studies are looking at ex pected job losses in the near future. The following table shows an over view of some of the different related studies between 2013 and 2017: Authors: Country/ Region Jobs at risk in percent (%) Based on Frey/ Osborne Own research Frey/ Osborne (2013)1 USA 47 X Bowles (2014)2 EU 46-62 X De Jong et al. (2014)3 Netherland 24 X Ekeland etal. (2015)4 Finnland, Nor way 33 X Brzeski et al. (2015)5 Germany 59 X Bonin et al. (2015)6 Germany 42 X Arntz et al. (2016)7 OECD 6-9 X World Bank (2016)8 World 37-83 X Chang/ Huynh (2016)9 ASEAN-5 44-70 X World Economic Forum (2016)10 15 countries - X Manyika et al. (2017)11 46 countries 5+ X Buhreretal. (2017)12 Germany 46 X 1 2 Frey/ Osborne (2013): The Future of Employment: How Susceptible are Jobs to Computerization?, University Oxford; Bowles 2014: Chart of the Week: 54% of EU jobs at risk of computerization, Brugel, Blog Post, online: computerisation/; ^ De Jon et al. (2014): De impact van 4 automatisering op de Nederiandse Arbeidsmafct. Een gedegen verkenning op basis van Data Analystics, Deioitte Research, Amstelveen; Ekeland et al. (2015): Computeriza tion threatens one-third of Finnish and Norwegian Employment, ETLA Briefs, 22. Research Institute ofthe Finnish Economy, Helsinki; ^ Brzeski et al. (2015): Die Roboter kommen. Folgen der Automatisierung fur den deutschen Arbeitsmarkt, ING DiBa Research, Frankfurt a.M; ^ Bonin at al. (2015): Obertragung der Studie von Frey/Osborne (2013) auf Deutschland, ZEW Kurzexpertise, Nr. 57, Mannheim; ^ Arntz et al. (2016): The Risk of Automation for Jobs in OECD Countries: A Comparative Analysis, OECD Social, Employfi 9 ment and Migration Working Papers, Nr. 189, OECD Publishing, Paris 2016; World Bank (2016): World Development Report2016, Washington DC; Chang/ Huynh (2016): 10ASEAN in Transformation. The Future of Jobs at Risk of Automation, International Labour Organisation Regional Office for Asia and the Pacific Working Paper, Nr. 9, Genf; 11World Economic Forum (2016): The Future of Jobs: Employment, Skills and Workforce Strategy for the Fourth Industrial Revolution; Manyika et al. (2017): A future that 12works: automation, employment, and productivity, McKinsey Global Institute; Buhreret al. (2017): The Effect of Digitalization on the LaborMarket, in: H. Ellermann; P. Kreutter, W. Messner (Hrso.l: The Palorave Handbook ofManaoino Continuous Business Transformation. Munchen. (similar: Heinen et al. 20171. As shown in the table, Frey and Osbornes (2013) are frequently quoted and their research paper from Oxford University “The future of em ployment” particularly, is a widely used research basis for other scholars. For this reason, this chapter will focus on the future of employment mostly based on the assumptions of Fred and Osborne's. The conse quences that will be addressed in the chapter are decisive for the further course of this research. They summarize the theoretical stance of the re searcher assume that massive job losses become a problem in the future. “Technology has always changed employment, but the rise o f ro botics and artificial intelligence could transform it beyond recognition“ (Oxford University 2016). 57 Frey and Osbornes proclaimed that AI and robots will replace approxi mately 47 percent of all US jobs over the next ten to twenty years (cf. Frey/ Osbome 2013: 38; Hummert et al. 2018). That is one result of their study, which examined more than 700 detailed forms of occupation, not ing the types of tasks workers performed and the skills required. By weighting these chosen factors, as well as the engineering obstacles cur rently preventing computerization, they assessed the degree to which these occupations may be automated in the near future. Particularly, the availability of big data was identified as one major trend that is given engineers a large dataset to work with. This has made it possible for computers to deal with problems that, until recently, only people could handle. The job cuts will be forced by robots and AI (cf. Dirican 2015; Sandhana 2013). Both also believe that these job losses will occur in two stages and are not limited to neither unskilled labor jobs nor blue-collar trades (cf. Ox ford University 2018; Lant 2017b). In the first stage, computers will start replacing people in especially vulnerable fields like transportation or lo gistics, production labor, and administrative support. Furthermore, jobs in sales, services or construction will too be lost. By the end of the first stage, they proclaim that the rate of replacement will slow down due to bottlenecks in harder-to-automate fields such as engineering. The first stage will be followed by a second stage, which includes another wave of computerization, dependent upon the development of strong artificial intelligence. This second stage could put more complex jobs for example in management, science and engineering, and the arts at risk (cf. Rutkin 2013). "Our findings thus imply that as technology races ahead, low-skill workers will reallocate to tasks that are non- susceptible to comput erization - i.e., tasks requiring creative and social intelligence. For workers to win the race, however, they will have to acquire creative and social skills." (Sandhana 2013). Frey and Osborne further argue that jobs that require social intelligence, creativity or cognition and fine motor skills will be among those activi ties that better secure against disruption of AI or robots. It is expected that occupations with these job profiles will become more important (cf. Eichhorst et al., 2015; U.S. Government 2016; Mahdawi 2017). So ac quiring more creative and social skills coupled with higher education level seems necessary, as high-skill jobs are less likely to be replaced anytime soon. 58 Many other experts have confirmed the main results of these studies. Economist of Harvard University, Larry Summers, for example, warns of the threat of massive job losses, whereby he warrants attention to the issue that there may be more industries that lose jobs than industries that create jobs (cf. Turner 2014). Furthermore, the World Economic Forum estimated that 65 percent of children entering primary schools now grow up to work in jobs that do not yet exist (cf. Macmillen 2017). Erik Brynjolfsson and Andrew McAfee from the Massachusetts Institute of Technology sum up this development in their book "The Second M a chine Age" in 2014. They underline the high risk of massive job cuts and the mass-unemployment as a transitional phenomenon as inevitable (cf. Brynjolfsson et al. 2014). From today's point of view, even a part of the new IT jobs introduced by digitization could be directly replaced by AI anytime soon. In conclusion, every technological transformation de stroys jobs, but also creates new ones. Yet, AI-driven automaton will be the first transformation that destroys more jobs than it will create new one. However, this rather pessimistic future scenario is also opposed by optimists who believe that AI-driven automation will not cause massivejob losses. “Some optimists argue that AI is no different than technologies that came before it and that centuries o f fears that machines will replace human labor have proven unfounded, with machines instead creat ing previously unimagined jobs and raising incomes.“ (Furman 2016). This argument is taken up again in the critical review (see Chapter 9) of this work but should not be considered in the context of further work. In the following course of this research, it should be assumed that the worse scenario describes the mass unemployment as a conceivable future sce nario. Due to AI-driven automation, also socio-economic inequality may increase sharply. Several scholars claim that even if technological auto mation may not increase unemployment, it can destroy middle-range jobs while increasing those on the low and high ends, augmenting social inequality as the pay between low and high-end jobs is amplified (cf. Makridakis 2017: 19ff; Maney 2016; Brynjolfsson et al. 2014; O'Neil 2016). Therefore, the focus should evolve more around the question of how much people will get paid if they will lose their jobs (cf. Kletzer 59 2018). As an interim result, possible future scenarios characterized by mass unemployment or even strong social-economic inequalities can be summarized. Nevertheless, there is uncertainty about these possible socio-economic consequences of AI, regarding exactly when and how they will occur in the future. According to Bill Gates talking about technology: “We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next 10.” (Wray 2017). Moreover, AI is a non-linear, tightly coupled and an astonishing tech nology, which makes its future even harder to predict (cf. D'Monte 2018; Metz 2018; Dempsey 2017). However, based on the research so far: Strong AI will change the information society including its economy and landscape of work. There is a consensus in the literature that AI will change the way humans work. The question is when and in which man ner this change happen. Overall, based on the concept of the information society and the analysis of the information economy, two possible scenarios should be considered further. Based on this, political options and possible policy outcomes will be discussed in the next chapter. The first scenario is oriented on the short- and mid-term: “Increasing human-machine integration and transformation” . The second one is more oriented on the mid- and long-term: “Idle human resources, further caused by a huge skill gap - even the highest skilled knowledge worker could end up with machines; discussing massive unemployment and ris ing inequality. 7.1. Short- Midterm Szenario: Increasing Human-Machine Integration and Transformation All in all, based on previous research, an increasing human-machine in tegration can be observed in most industries of the information economy. Furthermore, it is assumed that this integration will continue to increase and become deeper than ever. As mentioned above, this human-machine collaboration can go many different ways; constituting a great potential for the entire information economy. The following chapter will briefly describe this expected human-machine symbiosis. 60 The rise of AI already presents a shifting division of work between humans and machines. The goal is to bring their different strengths to gether. This means every part could concentrate on individual tasks per formed best by each party, respectively. For example, on the one hand, humans are good at creative thinking, problem-solving, conversations, generalization, and abstraction. On the other hand, machines are better for accessing databases, performing repetitive tasks, mathematical calcu lations and lifting heavy objects. So, by concentrating on their individual strengths, a human-machine collaboration can be very useful, which has already shown to lead to increased efficiency in the economy. Regarding the process o f decision making, for instance, the partnership between human decision makers and AI can play out in two ways: First, humans and AI technologies can collaborate to deal with different aspects of de cision-making. AI is likely to be well positioned to tackle complexity issues by using analytical approaches. This allows humans to focus more on uncertainty and equivocality, using more creative and intuitive ap proaches. Second, even the most complex decisions, in which AI has a competitive edge, are likely to entail elements of equivocality and uncer tainty. Therefore, humans continue to play an important role in almost all complex situations, as do AI in the face of equivocality and uncer tainty. Against this background, collaboration between humans and AI in a de cision-making situation can be illustrated using the following graphic based on complexity, ambiguity and uncertainty: Decide where to seek, and gather data. Choose among options with equal data support. Complexity Collect, curate, process, and analyze data.Artificial Intelligence + Negotiate, build consensus, and rally support. Equivocality Analyze se n tim e n ts, and re p re se nt d iverse in te rpre ta tio ns.Artificial Intelligence + Figure VII: Human-AI Collaboration in decision-making. Source: Own graphic, similar: Jarrahi 2018. The graphic shows that the pervasive visions of a partnership between humans and machines suggest that machines should take care of less creative mundane tasks, allowing humans to focus on more abstract 61 work (cf. Jarrahi 2018: 9f.). In practice, this could work in a very simple way, without consuming too much time or extra work. For example, each firm could bring an AI software in their conference room. Then the AI should be able to listen to the conversation in a business meeting while continually searching the internet for information that might be relevant, then serve it up when asked. So, it could support with knowledge of the outside world that the humans might not be aware of, which will lead to better decisions. By use-cases like this, scholars al ready believe that artificial intelligence will also improve human intelli gence (cf. Maney 2016; Carter et al. 2018). Human-Machine Symbiosis in decision-making is just one example of describing benefits from hu man-machine collaboration. All in all, based on the previous research a partnership like this can be expected, even if human intelligence and machine intelligence comprise an entirely different nature. The pattern recognition capabilities of big data and deep learning algorithms will assist humans to make new discoveries if yet requiring human intelligence to guide the AI devices as to what patterns to look for (cf. Braga 2017: 2f.). 7.2. Mid- Longterm Szenario: Discussing e.g. Massive Unemployment, Inequality and other Security Issues Based on the first part of chapter seven and the analysis of strong AI in the information economy (See Chapter 5) a mid- to long-term scenario can be characterized in the following way: Caused by strong AI, there will be idle human resources, further caused by a huge skill gap. Thus, even the most skilled among the knowledge workers may end up with machines doing their job, which could lead to massive unemployment. Regardless of whether or not this will actually happen, a scenario like this seems possible based on the conducted research so far. Thus, this pessimistic scenario could involve raising social and political instability, lost identity, crime or other security issues, which seems to be relevant to describe in the first part of this chapter. The impact of this scenario on the political macro-level and societies will be described in the last part of this chapter. It is assumed that its major challenges for today's infor mation societies can be expected. The negative potential of mass unem ployment could massively influence the central functions of today's in formation societies. Starting with the development of job cuts, which in effect leads to less purchasing power of citizens and thus reduction of 62 demand. Also, the concept of traditional wage labor and workers unions erodes and remains at most only for a small part of the population. With out work, on an individual level, many citizens could lose their identity or even suffer psychological problems because they always were used to work. All these aspects could trigger a change in values or even a loss of confidence in the state. Looking back at history, it can be learned that especially times of value changes, whether through technological inno vations or external aggressors, are troubled times for all kinds of rela tionships within a society. These upheavals are also accompanied by problems that come to the state, for example, lower tax revenues for the state as fewer people work, but rising expenditure caused by social measures, such as unemployment benefits, will inevitably lead to massive financial gaps in the state. How can a state help compensate for the needs of the unemployed people to pay for their living if they have no money for that? Answering and in vestigating such a question might be beyond the scope of this chapter, but serious problems based on this scenario seem realistic. Especially when the negative effects increase and the needs of the citizens, such as work or equal compensation, to maintain their standard of living is not met, this leads to serious threats to politics and society. Social or political instability, lost identity, poverty, riots, popu lism, crime or other security issues can potentially also be seen as a by product of developments in rising unemployment. This raises the follow ing questions: How many unemployed or social inequalities the infor mation society can tolerate in the long run and when it collapses? Are existing societies prepared for, or able to cope with, the challenges of increasing levels of strong AI, and thus possible increases in unemploy ment? If the pessimistic line of this chapter is continued, for example, today's capitalism in its current form would no longer be conceivable. An increasing number of scientists are addressing the issue of rethinking today's social and economic systems and the development of alternative systems. More commonly accompanying terms are: „The end of capital ism has begun“, “radical new economic system of the future“, „postcapitalism“ or even „decline of democracies^ All these scholars recommend, that governments have to reinvent existing economies and societies to keep up with accelerating technology (cf. Brynjolfsson et al: 2014; Confino 2014; Mason 2015; Helbing et al. 2017). „It’s time to start discuss ing what kind of society we should construct around a labor- light econ omy. How should the abundance of such an economy be shared? How can the tendency of modem capitalism to produce high levels of inequal 63

Chapter Preview



The ongoing and seemingly unstoppable digital transformation brings forth new options, opportunities but also challenges to individuals, organizations, companies and societies alike. Governments are alarmed, realizing the potential consequences on the workforce, while also being apparently helpless against uncontrollable and powerful digital players such as Google or Facebook. As Henry Wittke shows, recent breakthroughs in the field of machine learning increase the potential of Artificial Intelligence to disrupt the world’s largest industries. Wittke attempts to provide a basic framework of what constitutes AI as well as to assess its impact on the Information Economy. What happens in case of rising mass unemployment or social inequality? What will be the effect on labor as a value system for today’s societies? Could the entire notion of capitalism be questioned in the wake of AI? The book aims to draw conclusions and give recommendations to policymakers.