Big Data and Artificial Intelligence in Management. Disruptive Technologies as a success factor for decision-making

Table of contents

Foreword

List of abbreviations

1 Introduction

2 Basics
2.1 Current economic environment of German companies
2.2 Information and data
2.3 Digitization and Big Data
2.4 Artificial Intelligence
2.5 Other relevant aspects
2.6 Entrepreneurial Management 4.0

3 Analysis of decision-making in management
3.1 Key decision-making functions of management
3.2 Key factors in decision-making
3.3 Human decision-maker versus artificial intelligence
3.4 Decision-making insights

4 Analysis of the feasibility of artificial intelligence in practice
4.1 Basic needs
4.2 Risks
4.3 Insights into feasibility

5 Concluding remarks

Bibliography

Foreword

In cult series such as Knight Rider or Hollywood films such as Star Wars the topic of artificial intelligence has been represented for a long time. But especially in recent years, the topic has made the leap from an artistically designed fiction to theoretical and applied science. I perceived the increasing presence in the trade and general press as an indication of the opportunities and risks of this new technology, which may fundamentally change entire industries.

Specialist articles and study results of my training partner made it clear to me the acute need for well-founded insights into the implication possibilities in the economy and their effectiveness in management decisions.

As part of my course of study "Accounting, Taxation, Business Law – Auditing" at the Baden-Württemberg Cooperative State University, I learned basic knowledge about the management of companies. With this work, I would like to scientifically substantiate my basic knowledge acquired in the lectures and consolidated in practice and relate it to the topic of artificial intelligence.

The aim of this work should not only be a scientifically sound commentary on the problem, but also the acquisition of specialist knowledge, which I can apply in my professional activities.

List of abbreviations

Abbreviation Meaning

AI Artificial Intelligence (engl. zu dt. "Artifizielle Intelligenz")

ARE Augmented Reality (engl. zu dt. "Erweiterte Realität")

BSI Federal Office for Information Security

DAX German Stock Index

GDPR General Data Protection Regulation

CEO Chief Executive Officer

CFO Chief Financial Officer

COO Chief Operating Officer

CDO Chief Digital Officer

CAIO Chief Artificial Intelligence Officer

E-commerce Electronic commerce, or online trading

e.V. Registered association

EY Ernst & Young

IT Information Technology

Iot Internet of Things

AI Artificial Intelligence

LKW Truck

MDAX Mid-Cap Dax

Mio. Millions

MIT Massachusetts Institute of Technology

Mrd. Billions

NIST National Institute of Standards and Technology

PC Personal Computer

PKW Automobile

PWC PricewaterhouseCoopers

RFID Radio-Frequency Identification

RPA Robotic-Process-Automation Robot-controlled process automation

SDAX Small-Cap Dax

SMS Short Message Service

US United States

UNITED STATES United States of America

US -Dollar United States Dollar

VDI Association of German Engineers

1 Introduction

From the founding of the first company to the present day, it has been the task of the management to make decisions that affect success. The demands on decision-makers have continued to increase over time.

Geopolitically, globally and technologically, the world is in a state of upheaval. Technical progress produces ever shorter innovation cycles and the processes and projects in companies are constantly gaining in complexity. Furthermore, the globalized economy is creating increasingly dynamic markets. New markets often open up just as quickly as they close or change again. Product portfolios are also becoming increasingly diversified. In addition, in times of online trading, sales channels are becoming more and more versatile. In a time that is characterized by the networking of the real with the virtual world, data emerges as the resource of the future. The huge amounts of data, also known as big data, demand their effective use.

The consequences for management are increasingly frequent, increasingly difficult and complex decisions. The multitude of possible solutions is also constantly expanding. Despite these increasing demands on management, the decision-making body remains constant. Since the beginning of entrepreneurship, people have been behind the decisions.

But the working time of this decision-maker is scarce and the question arises as to whether human capacity, intelligence and creativity can cope with the increasingly demanding decisions at all or whether there is a better decision-maker here. Currently, the focus is on artificial intelligence (AI). The presence of artificial intelligence is reflected, among other things, in the fact that the Award for Entrepreneur of the Year 2019 (International), awarded by the renowned company Ernst & Young (EY), was awarded to an entrepreneur specializing in AI industrial applications.

The question arises as to whether artificial intelligence, using big data, could act as a success factor in corporate decision-making?

The aim of this work should therefore be to comment on this question in a scientifically sound manner. To do this, it must first be recorded by what the success is measured.

Success is measured by whether an economic advantage is achieved from the point of view of the company as a whole. So either through a more efficient decision-making process or through more effective decisions. More effective decisions mean that the solution chosen by one decision-maker brings greater economic success to another. For example, existing value creation processes are optimized or new ones are developed. Through this definition, the quality of the decisions can be evaluated transparently and comprehensibly.

It is particularly noteworthy that there is a clear difference between the evaluation of a decision and a decision-maker. Because an objectively correct decision does not necessarily correlate with the competencies of a decision-maker. A common error of thought that can also have negative effects in decision-making. This risk is specifically addressed in Chapter 3.3.3. The differentiation takes place as follows:

Quality of the decision: For a statement about the quality of a decision, an objective perspective must be taken. Under the assumption of an omniscient expert, the decision is to be chosen as the best, which actually brings the greatest economic success taking into account all conceivable solution variants.

Quality of the decision-maker: For a statement about the quality of a decision-maker, a subjective perspective must be taken. From the point of view of the decision-maker and with his level of knowledge at the time of the decision, the decision that is most promising should be made.

It has already been made clear that decision-makers should not be judged by the result, but by the process. For this work, however, the two decision-makers humans and artificial intelligence are nevertheless judged at the most objectively considered economic result. The background is that in a subjective view, the focus is initially on the decision-maker at the moment of the decision. However, this situational scope of knowledge is highly differentiated between AI and humans and the elaboration of this is too extensive. Rather, it is also irrelevant from the point of view of the company whether subject 1 or 2 has made the best decision to be made from his point of view. As already mentioned, it is about what is objectively the economically better decision for the company.

Since the topics of management and artificial intelligence have a very far-reaching scope, their synthesis in this work often makes a distinction to aspects that can not be discussed further.

The substantive procedure can be briefly outlined as follows: First, the current economic environment in which companies in Germany are located is presented in the relevant aspects. The central technologies related to AI will also be presented. The topic includes a variety of new terms. Especially in the first part of the work, it should therefore come to the clarification of the relevant terms. This work deals with the decision-making of the management, for this reason the impact of the current entrepreneurial environment on the management is clarified. A variety of factors work into the decision-making process, which are classified as essential for this work and are therefore presented. This is followed by the investigation of how an AI deals with these factors compared to a human and which advantages or disadvantages are provided in each case.

After theoretically working out how an artificial intelligence performs with regard to an economic goal in the decision-making towards a human being, the practical feasibility is to be illuminated. The basic prerequisites of an AI are presented and possible risks associated with the implementation of AI are shown. After it has been worked out in theory who - from an economic point of view - could be the better decision-maker in today's world and the implementation has been examined for its practical feasibility, the problem can then be commented on in a well-founded manner.

This work claims to be a scientific work that largely refers to specialist literature from various university libraries. Developments on this topic are progressing on a daily basis. Interest from international research, business and politics is increasing. As a result, many current professional articles and other contributions by established experts are accessed via the Internet. Studies, surveys and comments from renowned consulting and auditing companies can also contribute insights. Panel discussions at universities also provide up-to-date opinions from politics and science.

The topicality of the topic requires, as can be seen in the bibliography, sources essentially from the last two to three years. Decision-making scenarios against an economic background require reflection with established models and approaches for economic soundness, and so sources up to the year 1637 can also be found in the bibliography.

2 Basics

2.1 Current economic environment of German companies

The corporate landscape in Germany is characterized by industrial companies, with the manufacturing sector alone generating sales of 2,169.76 billion euros in 2017.¹ German industry is in the age known as Industry 4.0.² This phase, which will be discussed in more detail in the course, is the result of three serious industrial revolutions. The following is a brief outline of these stages of development:

Industry 1.0: The first industrial revolution is essentially associated with the use of large machines, such as the loom.³ In addition, it was characterized by a production supported by steam or hydroelectric engines.⁴ Production was predominantly regional, which is why individual market fluctuations had a strong impact on the business and thus also directly on the labour supply. The resulting impoverishment of the working class resulted in extensive protest movements.⁵

Industry 2.0: The second industrial revolution is characterized by the transition to mass production, in which the guiding principle of a more economical use of resources was in the foreground.⁶ Among other things, the concept of assembly line work was established.⁷

Industry 3.0: The third industrial revolution resulted from the use of information technologies and the first automations in production processes.⁸ An attempt was made to replace human labour power more and more.⁹

Industry 4.0: Characteristic of the fourth industrial revolution is digitization and Internet-based communication, which is also used within production. The networking of production lines within a factory opens up new opportunities to increase profitability.¹⁰ The guiding principle of this time is the networking of the real with the virtual world.¹¹

As part of the global economy, Germany is experiencing a time that seems to be as diverse, as dynamic and as complex as never before. A glance at the press, be it trade journals such as the Harvard Business Magazine or even daily newspapers such as the Frankfurter Allgemeine Zeitung, increasingly shows companies that have a worldwide network through group structures and business partners. It also reports on a time of digitization and automation. Technical progress produces highly developed technologies. The technologies known as "disruptive technologies" could fundamentally change the way an industry works.¹² E-commerce, i.e. sales via the most diverse channels of the Internet, has long since ceased to be just a trend, but an integral part of retail. In addition to international online retailers such as Amazon or Alibaba, German family businesses such as the Otto Group also show that the distribution channel has established itself in the online retail sector with a turnover of 5.4 billion euros in the 2017/18 financial year.¹³

Innovation cycles are shortening, and new products are coming onto the market almost every day. In the past, it was primarily about the introduction of a new product, such as a light bulb, a refrigerator or a television. There is currently a time when it seems necessary to launch a new, optimized product on the market every year or every six months. For example, the iPhone is already in its 14th generation. In addition, customers today expect a variety of product variants. While there used to be a variant of the iPhone, one generation currently includes three variants (iPhone XR, iPhone XS and iPhone XS Max).¹⁴ A look at the offer of a large automotive group also supports this trend. In the passenger car segment alone, Mercedes-Benz's range of products developed from an initial six to 33 models offered in 2018 (purely Mercedes-Benz vehicles without Smart).¹⁵

In summary, companies are in a challenging, highly versatile and dynamic environment. Digitalization and networking are constantly spreading.

2.2 Information and data

Chapter 2.3 will give an insight into the topics of "digitization" and "big data". For dealing with these terms, a certain basic knowledge is a prerequisite, which should be conveyed in advance.

The terms "data" and "information" are essential here, first of all to clarify them. In the introduction, it should be emphasized that the terms are usually used as synonyms in everyday life and also in science an equation of the terms can be recognized. Nevertheless, there are various approaches to differentiation, but especially the concept of information itself is also attributed a large number of understandings. For this work, it is therefore important to commit oneself to an understanding and to always observe the local understanding of the term when using sources.

The term "information" sometimes experiences physical understandings, such as information, as a pattern of the organization of matter and energy. Furthermore, there are various approaches of a semiotic nature, essentially an objective understanding, in the sense of information as a true fact, which exists independently of humans. On the other hand, a subjective understanding in the sense of information as individually varying depending on the recipient. Furthermore, an intersubjective understanding, here information is considered dependent on the social group in which it occurs.¹⁶ The disambiguation thus has a high degree of complexity, the scope of which is not proportional to the core of this work. Therefore, this is no longer the subject of the matter and the terms for this work are based on the following understandings. These are coordinated in such a way that they meet the most common understandings in the sources used.

According to the definition of the Gabler Business Dictionary, information is defined as the new part of a message for the recipient.¹⁷ Data, even if they are often equated with the concept of information in the literature, are set differently from this. Objective and measurable facts are assumed as data. Data is thus understood as a kind of precursor to information, as objective data that can then be interpreted in a subject-oriented manner and thus become information at the respective recipient.

2.3 Digitization and Big Data

2.3.1 Digitization

The term digitization is essentially used under two meanings. On the one hand, in the sense of transforming information and communications from the real to the virtual, digital world.¹⁸ For example, the company Ernst & Young has switched from paper-based documentation to digitalization in the audit of annual financial statements. Where in the past the examiners documented examination procedures by hand, now it is mainly digitally documented in the computer. Digitization is not only a concept of the economy, digitization is a far-reaching topic. As an example from everyday life, the transition from letter communication to communication via messaging services such as "SMS" or "WhatsApp" can serve.

The second meaning of the term arises from the fact that digitization is omnipresent and significantly changes so many processes. Digitization is therefore understood as a digital revolution.¹⁹

Digitization primarily generates the association with the optimization of processes, which are intended to provide an efficiency and effectiveness advantage in digitized form. As part of a panel discussion at the University of Stuttgart under the question "Let's make the transition to Silicon Ländle", former Telekom board member and member of the Bundestag Thomas Sattelberger pointed out that an opportunity of digitization is often missed. He explained that digitization is not only about digitizing existing value-added processes and thus optimizing them, but also about leveraging new value creation processes. As an example, he cites the Daimler Group, which no longer calls itself a car manufacturer, but calls itself a mobility service provider. And through services such as car sharing (Car2Go), new, in the form of a sharing economy company, also beyond the optimization of its previous value creation processes²⁰ creates. However, compared to the USA and other nations, this is not yet recognized enough, according to Sattelberger.²¹

The entrepreneur and investor in the IT segment Marc Andreessen made a statement in the Wall Street Journal in 2011 that metaphorically sums up digitization in one sentence to this day. He put in the room: "Software is eating the world", translated into German "Software eats up the world".²² The sentence essentially reveals two statements, on the one hand digitization embraces the whole world, which can be confirmed, because in fact there are almost no areas that have no points of contact with digitization. On the other hand, the sentence reflects people's fears. Because it is feared by some, presumably initiated by science fiction films, that the software "eats" people, in the sense of gaining control over them.

Internet and software companies have corporate values in 2019 in some four-digit billion ranges: Facebook $537.99 billion, Google/Alphabet $766.66 billion and Microsoft even at $1,036 billion.²³ This shows the significance of software already. It is particularly interesting to look at the following two companies. Uber, a transportation company, was worth $76 billion in 2018, but without owning its own transportation vehicles. AirBnB, a company that provides overnight accommodation, was worth $31 billion in 2018, without owning its own overnight accommodations.²⁴ These two companies thus illustrate the possibilities that software reveals, because both companies do not own the objects that are actually the focus, but only an intermediary software platform.

Digitization can thus be described as in full swing and is fundamentally changing not only private everyday life and company processes but also a large number of other areas.

2.3.2 Big Data

In the course of digitization, the term big data came up. First of all, this simply means a large amount of data. However, when talking about big data in science or business, big data has four dimensions. These four dimensions are: Volume, Velocity, Variety and Veracity. Volume describes the amount of data. Velocity describes the speed at which the data is collected, but also the speed at which it is stored and processed. Variety describes the diversification of data, i.e. the mix of structured and unstructured data. Veracity describes the challenges associated with the previous three dimensions, such as the trustworthiness of the data or the quality.²⁵ Big data thus brings together these four central dimensions for the use of data in one term. Big data, for example, commonly stands for the emergence of a large amount of structured and unstructured data that can be collected and processed at high speed. This definition will be used in the further course of the work.

In the age of Industry 4.0, in addition to natural resources such as manpower and raw materials, data can also be regarded as a resource. Data has a significant difference from natural resources, it is not scarce.²⁶ The interest in them is particularly increased by the ever-increasing speed at which data can be collected and processed.²⁷ The amount of data continues to increase steadily and seems to have no limit.²⁸ According to a statement by entrepreneur and Professor Peter Gentsch in 2018, the amount of available data doubles every two years.²⁹ The reason why the amount of available data has accumulated so rapidly in recent years includes the expansion of the Internet and its use, as well as more and more built-in sensors.³⁰ Among other things, roads, cars or houses are becoming more and more networked. And the widespread use of cashless payment options from credit cards to mobile payment services such as Apple Pay, from a major corporation of the same name, also increases the data collection potential. The amount of data that can be collected is strongly related to the number of users of a technology or software that can collect data. A study by EY shows that the time it takes for a technology to attract users is getting shorter and shorter. The telephone invented in 1923 took about 75 years to reach 50 million users, the TV in 1970 only 13 years, the Internet in 2004 only 4 years and, for example, the game Angry Birds Space needed only 35 days in 2012 to record 50 million users worldwide.³¹

According to an EY study, it is predicted that around 50 billion devices will be connected to the Internet by 2020. This also includes RFID chips³² and sensors.³³ An EY survey of German SMEs gave an insight into how networked production already is today. In the representative survey "Industry 4.0 in German SMEs" by EY, in which 1,157 medium-sized industrial companies in Germany were surveyed, only 23 percent stated that they had not yet digitally networked production and did not plan to do so.³⁴

Big data can become smart data through artificial intelligence, i.e. data that can be used effectively.³⁵ Just as AI is a key function for the usability of big data, it is also inverse. Because before an AI can work effectively, it must go through an intensive learning phase. Big data can help provide training data here. The more extensive these are, the better the later results are usually.³⁶ The following example shows how large the data volume should be. It is assumed that an AI should assign existing images to given categories and that the learning phase takes place on the basis of supervised deep learning, which is discussed in more detail in Chapter 2.4. The requirements then include 5,000 sample images per category for reference, as well as 10 million test images used to perform the supervised training.³⁷ Since so much data is required for such a comparatively simple task, it becomes clear what key role big data can play for AI, especially in complex tasks.

2.4 Artificial Intelligence

2.4.1 Aspects of the history of origin

This thesis deals with the topic of artificial intelligence under an economic background. However, before AI research is outlined in time and the term is explained in its relevant meaning, the motif AI in film and literature should be addressed. The broad use of this motif could be seen as an essential driver for research. Thus, the fictional ideas in film and literature could have inspired to deal with the real feasibility of such ideas.

According to a study by the Gesellschaft für Informatik e.V., the best-known artificial intelligences from film and literature in Germany are terminator, from the film of the same name, R2-D2, from the Star Wars series and K.I.T.T. from the Knight Rider series.³⁸ In addition to these, however, there are a large number of other films and books, especially internationally, in which AI is a leitmotif, current examples are transcendence or Ex Machina. The drama published in 1920 Rossum's Universal Robots, is certainly one of the first pieces in which robots threaten to take over the world.

[...]

---

¹ cf. Statistisches Bundesamt (Federal Statistical Office) (2018)

² cf. Hermeier, B. u.a. (2018), p. 4.

³ cf. Hirsch-Kreinsen, H. / Minssen, H. (2017), p.176

⁴ cf. Hermeier, B. u.a. (2019), P.112

⁵ cf. Hirsch-Kreinsen, H. / Minssen, H. (2017), p.176

⁶ cf. Ibid., p.176

⁷ cf. Hermeier, B. u.a. (2019), P.112

⁸ cf. Hirsch-Kreinsen, H. / Minssen, H. (2017), p.177

⁹ cf. Hermeier, B. u.a. (2019), P.122

¹⁰ cf. Pfrommer, J. u.a. (2015), p. 379.

¹¹ cf. Hermeier, B. u.a. (2019), P.112

¹² cf. Cambridge University Press (2013)

¹³ cf. Otto Group (2019)

¹⁴ cf. Chip.de (undated)

¹⁵ cf. Daimler (2018)

¹⁶ cf. Boel, S. / Cecez-Kecmanovic, D. (2011), p. 3 ff.

¹⁷ cf. Lackes, R. u.a. (2018)

¹⁸ cf. Bendel, O. (undated)

¹⁹ cf. Bendel, O. (undated)

²⁰ Sharing economy stands for the sector of collective use of assets, cf. Haese, M. (2015), p.1

²¹ cf. University of Stuttgart (2019)

²² Andreessen, M. (2011)

²³ cf. Yahoo (2019)

²⁴ cf. Bloomberg (2019)

²⁵ cf. Gentsch, P. (2018), p.9 f.

²⁶ cf. Oliver, F. (2019), p.21

²⁷ cf. Gentsch, P. (2018), p.7

²⁸ cf. Oliver, F. (2019), p.21

²⁹ cf. Gentsch, P. (2018), p.13

³⁰ cf. Ibid., p.7

³¹ cf. Ernst & Young AG (2017), p.10

³² RFID (Radio Frequency Identification) is a technology that makes it possible to make contactless identifications, e.g. key/access cards, cf. Krieger, W. (2018)

³³ cf. Ernst & Young AG (2017), p.10

³⁴ cf. Ernst & Young GmbH (2018a)

³⁵ cf. Schneider / Vöpel / Weis (2018), p.8

³⁶ cf. Gentsch, P. (2018), p.11

³⁷ cf. Gentsch, P. (2018), p.11

³⁸ cf. Gesellschaft für Informatik e.V. (o.J.)