COMPANIES TODAY HAVE become relentless data-processing machines. It is therefore no coincidence that the notion of big data has become such a buzzword in the business community. Companies have realized that they have so much in-house data available, which, if processed and analyzed properly, could drive revenue and make them more effective. Business strategies today then need to focus on data management in combination with evidence-based execution.
It is becoming clear that for contemporary organizations, data is a sacred component of performance management. In fact, data is increasingly being referred to as the new oil.89,90 This increased focus on data is something that all of us are confronted with on a daily basis. Organizations require as many personal details as possible, communicate a wide range of targets to achieve and set out to assess those targets in a timely fashion. As a result, continuous evaluations and additional requests for supplemental data have become a routine process, revealing large data sets for each employee. The challenge for any smart company today then is to manage and connect all these data sets to optimize execution and enhance performance.
This reality has implications for our jobs today. First of all, due to an increased focus on acquiring data, most of us are becoming occupied with managing and updating our own data files (as often requested by the HR department) and documenting our activities. Second, all this data needs to be submitted on a regular basis to internal systems in the organization and will be used by our managers to evaluate our performance. This way of working obviously generates a lot of data, but most companies, eager to call themselves skilled in management of big data, consider this strategy essential to their reputation of being a smart company.
On the other hand, the entire process of managing and submitting data, which is then continuously evaluated and updated, also takes a lot of effort and time. This contributes significantly to additional layers of bureaucracy. Indeed, the fact that the focus on big data promotes bureaucracy in organizations, combined with the observation that companies are not necessarily good at using big data in effective ways, makes one wonder whether the whole obsession with data collection, evaluation and use is actually worth all the effort.
Presumably there must be an upside to this way of working. And there is, if we consider well-informed management as a key to an organization’s health and prosperity. We all know that good performance is key for companies to compete and to survive in the long term. This managerial truth requires organizations to engage in performance management strategies rooted in updating, processing and evaluating the data of each employee. This way of working allows for managers to give feedback to employees in a data-driven way, which is believed to be more accurate and should therefore be more effective in making each employee perform better. Today’s challenge for managers is thus to promote performance by making the best use of the data available. If management’s aim is to promote stability, order and consistency in producing the best outcomes possible, then working with data is a key principle to optimize the management of any organization.
Do I need to become a coder?
When you read all of this, you may consider the thought that these data-driven processes, ultimately leading to more optimized performance management, run in very rational and consistent (even replicable) ways. If this is the case, a question that emerges rapidly is whether we really need humans to execute the data driven management aspects of an organization? In fact, if you are asking this question right now, you may have touched upon one of the most important questions in today’s automation age: should management as we know it become a job performed by algorithms?
If the aim is to use data in more optimal ways to rationally and consistently empower better performance management, then humans may not be fit for purpose, especially not if the amount of data that needs to be dealt with is continuously increasing. To answer this question, it is useful to first look at the exact qualities attributed to algorithms and how they can help organizations run in more successful ways.
The one thing that makes algorithms uniquely fitted to any job related to data management is that they work so much faster than humans in collecting, evaluating and integrating data. Second, they work more systematically and are rational in arriving at conclusions (no sentiments or intentions are at play), making their predictions more impartial and accurate. And, finally, employing algorithms is way cheaper than hiring humans for the same job.
Putting all these qualities together, and assuming that any organization wants to drive down costs and raise up the effectiveness of execution, algorithms do appear to be a better substitute for human managers. Tasks will be done faster, at a lower cost and with a higher level of accuracy. It sounds like the ideal scenario for any business leader working hard to optimize the working of their company. In fact, in light of how management is defined in the literature (see the previous chapter) and the current trend to narrow down managerial work to mainly administration and performance management, we, as humans, seem to have put ourselves in a position where we can be easily replaced by algorithms. It should therefore be no surprise that voices exist saying that the end of human managers is in sight.
Algorithms clearly have the winning hand when it comes down to being the rational administrator for which we have long searched. In addition, because of their fast and consistent way of working, they are unbeatable in providing the most optimal use of any data available. It also excels in the practice of evaluating and even categorizing the performance of any employee. This reality has not escaped the attention of our managers today. In the last few years, I have noticed that in my (E)MBA classes, students almost immediately start sweating when they discuss the future of work and the roles of humans and algorithms. As I explained earlier, MBA stands for master of business administration. In theory, this categorizes any MBA student as an administrator trained to run companies in a systematic way, with the belief that management is based on rationality as a guideline. It is then no surprise that once these students realize that algorithms are designed to be much better in this kind of rational way of working, many of them worry about their future.
What is the value of being educated in rational business strategies by business schools if the underlying assumptions of those strategies can be better executed by an algorithm? Wouldn’t it be a better investment for them to become a coder and leave the business wisdom to algorithms? After all, they could design an algorithm and be entirely in charge of it, rather than competing against the algorithm’s rationality?
Does human intuition pay off?
Don’t worry, as I explain later, there is no need to become a coder. There does, however, exist a need to train our business students in new ways, so that they are better informed on the application of algorithms and what it might mean for today’s management. They also need to learn how to deal with this trend towards automated management by knowing which skills to focus on in the future. But, before I do that, let us be fair and see whether there are really no objections towards using algorithms as the perfect substitute for the human manager.
The initial response could be that as machine intelligence improves (and it is doing so almost on a daily basis), the more likely it is that humans will lose this debate. So, rather than focusing on whether humans can withstand the rise of algorithms in management positions, wouldn’t it be better to shift our focus on the unique human abilities that algorithms do not possess. This revised focus may help us to get a better picture of how management will change by means of automation as well as what role humans will play.
So, what traits can be considered uniquely human? Interviewing managers who have worked in recruitment for several decades, many point out to me the importance of human intuition. Some of them have developed a kind of sixth sense that when looking at a CV they immediately feel whether this person is suitable or not. For those who have been in the recruitment industry a long time, their predictions are usually pretty accurate! Since it is essential for a manager to ensure good performance, they need to be able to bring in the right kind of people. Indeed, recruiting and finding the right fit between the new hire and the organization is an important management responsibility. And, if human intuition seems to be important in this, could it then be that the human side of management still has a shot at surviving? But, before we answer that question, let us first try to understand exactly what human intuition is.
Is human intuition really such a mysterious feeling? Unfortunately, to those who like to believe so, behavioral science has made clear that intuitive decision-making appears to be the result of training, although a kind of training of which one is often not aware. How? Well, there is a reason why I emphasized that the intuition of long-term recruiters is often accurate in predicting positive hiring decisions. Day in, day out, for a very long time, they have been doing the same task. As a result, they have so much experience with all kinds of situations that their brain is trained to know immediately what to do and what to expect. Over the years, a recruiter will build up, both consciously and unconsciously, a multitude of relevant information. And so, many years down the line, a recruiter’s intuition will have become like a fine-tuned instrument – an almost perfect recruiting tool.
However, recruiters who have been in the job for only a short time may know the requirements of the job and what information to look at, but their intuition is unlikely to make them successful. This is because they did not have time yet to train their brain in unconscious ways to make the best decisions possible. Furthermore, even for those very experienced recruiters, the positive influence of their intuition is still limited. Indeed, simply ask these very experienced recruiters to make a quick decision in an area different from their own expertise and their intuition will not be that effective.
The human sense of intuition as a powerful tool to optimize predictions is thus very much expertise related and takes a long time to perfect. Importantly, for those non-recruiters reading this book, looking at intuition as a subconscious process works in similar ways across any job in business.
My favorite example to demonstrate this assumption is to refer to the workings of company boards. I always ask my students which team in their company is the one that makes decisions almost always in intuitive ways. Usually, I get responses like the sales team, R&D, customer management and so forth. When I mention the board of directors, some look up in surprise while others cannot suppress their laughter, as if they are immediately recalling board decisions that were flawed by what we call human intuition. Is this such a surprise?
If we look closer at how boards usually make decisions, it is interesting to observe that in the decision-making process many board members overly rely on their sense of intuition. Many of them even explicitly say that their gut feeling is the advisor they trust the most to make important decisions. The reasoning behind this is always that they have the idea that their intuition helped them to get to the position they are in now, so, it is the best predictor for making good decisions. As a result, if something does not feel good, they vote against it, but if it feels good, they vote in favor. The reason for them to rely so much on their intuition is simple. Across the many management positions they have held over the years, they have acquired experience in all kind of business decisions. They have experienced and seen the outcomes of those many business decisions and have internally built an image and feeling that they have seen it all.
The brain plays an important role here. In a way, their brain has – on a subconscious level – stored all this experience, all the data, acquired during the course of their career. So when, after many years as board members, they are asked to make the ultimate decision, the stored information in their brain speaks to them via the magical process of intuition. This is the reason why their gut so often talks to them. And, because they literally feel it (or, at least, in an imagined way), they also trust it and consider it to be true.
The human brain shines, the algorithm gets things done
So, the process of intuition teaches us that humans can actually become pretty reliable and accurate data-processing machines. And they can use their acquired knowledge as input for the decision-making process. If this is the case, then why bother to automate management? Isn’t it the case that what algorithms do today is essentially the same process?
Algorithms quickly assess, analyze and connect the different pieces of data and start self-learning (just like humans subconsciously process experiences and learn from them). The way algorithms work ultimately enables them to identify trends that help bring transparency to complexity and increase the accuracy of predictions. The rapid advancement of AI is based on our increased insight into how the human brain works. As such, it is clear that algorithms are modelled on how we think the human brain approaches and analyzes information. The only big difference seems to be that algorithms work so much faster – humans clearly need many years for their intuitive sense of accuracy to kick in. Given the fact that algorithms are modelled after the human brain, it should also not be a surprise that we are thinking of applying algorithms to the work performed by humans. In organizations, this will first and foremost include administration and performance management.
Before we move on to see how grandiose and beautiful the algorithmic side of management may turn out to be, I would quickly like to note that by drawing a parallel between the human brain and algorithms, I am not claiming that the human brain is something that is not advanced. In fact, as we strive to replicate the human brain, people may even get the impression that the organ is not such a big thing after all. Well, let me make this clear to you: the human brain is a big thing!
If you talk to any neuroscientist, psychologist, or brain scientist, they will all tell you in the most beautiful way just how magical and complex the workings of the human brain are. And the reality is that today, in this respect, our technological advancements are just not that advanced. This is especially apparent when it comes to the potential for technology to truly understand how human intelligence operates.91 To expand further, the human brain is so complex that we actually still know very little about it. For example, we still do not fully understand how the electrical signals of neurons influence the brain’s functioning, something which is key to the learning and associative-thinking process. Besides this, our methods to examine all of these internal brain functions are far from perfect.92
Although we know far less of the human brain than we would like, technology develops quickly and the benefits of applying it are growing every day. As such, these advances force us to be serious in thinking about automating management. Even though we are not yet able to develop algorithms that capture the complexity of the human brain, many managerial tasks can actually be done by algorithms in much faster and more accurate ways. Many companies work in volatile and fast moving environments. Under such circumstances, most companies do not have the luxury to train humans to become the perfect manager. They do not have the resources or the time to employ managers for so many years, just hoping that at the end of the day they will deliver the perfect result.
And why should they? Just think about it. If you have algorithms ready to jump in to perform the basic management functions in more optimal ways, why hesitate? Take, for example, the management of recruiting. In this field, evidence is mounting that algorithms are already producing better results – despite not having the sophisticated intuition of experienced human managers available. The paper from the National Bureau of Economic Research that I referred to earlier corroborates this idea. This paper examined the employment record of 300,000 low-skill, service sector workers across 15 companies. In this sector, workers usually leave their job quickly (they stay for an average of 99 days). Interestingly, however, when an algorithm was involved in hiring these workers, they stayed in the job 15% longer.93 Findings like this clearly indicate that algorithms can bring immediate added value to job performance.
Obviously, the enormous cost benefits and immediate value created has not escaped the attention of the business world either. It is therefore no surprise that today we see strong signs that the traditional management job is under threat from algorithms and there is a global trend towards automating management functions. This is especially true for managerial functions that include a focus on data input and maintenance. Processing is also expected to be automated in the near future.
Data management is needed to facilitate the use of the correct type of data to co-ordinate projects effectively. Performance should then increase accordingly. Interestingly, surveys reveal that on average, managers spend 54% of their time on exactly those kind of administrative tasks.94 As algorithms are the real experts when it comes down to working with data, automation of these functions is taken for granted. In fact, most business leaders believe that it is a given that administrative jobs will face massive replacement by algorithms. This is not simply a prediction for the future, it is something that is already happening today. Examples abound: IBM, for example, applies the algorithm Watson Talent to its own HR teams to promote speed, efficiency and the optimal use of their operations.95
Another example of such automation is the use of Robotic Process Automation (RPA). RPA uses software algorithms to closely replicate repetitive tasks like moving data between two spreadsheets. And, finally, especially within the context of HR management, the employment of algorithms to conduct repetitive administrative tasks has already been proven to be effective. For example, every time a company hires a new employee, algorithms can automatically update data files, including vacancies; create accounts for the new employee within the employee system; and integrate the software systems so the data of the new recruit can be accessed by different departments across the company. All of these examples show that today the automation of administrative tasks is mainly focused on the facilitation of repetitive and routine operational decision processes.96,97,98,99,100,101
Bring out the blockchain
OK, algorithms are executing managerial responsibilities, but that does not make them managers. Does it? Maybe algorithms are not real managers yet, but what all these examples demonstrate is that companies are not just thinking about building a work culture of automated management, they are already doing it.
Increasing amounts of evidence suggests that the practice of automation will not only be limited to simple administrative tasks that can be executed in efficient and fast ways. In fact, the trend to automate management roles will gradually involve more complex jobs, such as evaluating human employees to facilitate more effective collaborations and exchanges at work. Algorithms will not only be used to document and analyze personal data, but also to monitor employees by collecting new data that could complement the data collected when those employees were recruited. For example, algorithms are already monitoring the time you spend on the internet, your whereabouts within the company and even your health data.102
Yes, algorithms seem to be assuming the role of big brother and, as such, seem to become the manager that we know today. The contemporary manager evaluates and monitors your progress and performance. In light of this movement of algorithms into a role of managing (some may say monitoring) others, an interesting application in this area concerns the potential employment of blockchain technology to manage work relations in organizations.
Blockchain is mostly known as the underlying technology of applications, such as Bitcoin. At the same time, the technology is also increasingly being seen as a way of changing how our companies work. One specific application that was pointed out in a recent Deloitte survey is to use this technology to perform the basic functions of management as well as motivating employees more specifically.103 This survey revealed that among 1,386 senior executives in 12 nations, 83% saw compelling ways for blockchain to be used by their organizations. Of specific interest to our discussion here is the finding that 86% of these executives believe that blockchain technology can be applied to the aspects of management that involve leading an organization. How could blockchain technology be used to achieve this goal?
Let us first consider, what is blockchain? Blockchain is a distributed database composed of a chain of blocks in chronological sequence. Each block is a collection of data. Blockchain records data about past behaviors of individuals who are all participating within the same interconnected network. The technology behind blockchain thus builds a platform that holds a transparent and immutable record of past events relevant to all individuals within a shared network. This network could be a team, department or even the organization in its entirety. Transparency is created based on the history of interactions taking place within the network, which means that a blockchain should have the ability to make individual parties trust each other. Indeed, in 2015, The Economist put blockchain on its cover and called it “the trust machine”. Because of its supposed ability to create trust, it should be possible for blockchain to manage behavior within companies.
The key element in blockchain’s supposed role as the trust builder is that it creates a risk-free environment. The past of every individual involved in the network is controlled and the risk of exploitation is therefore virtually zero. When it then comes down to managing an organization, blockchain technology is regarded as suitable for delivering assurances (a stamp of verification that interactions are safe) that can increase co-operation, while protecting individual employees’ interests from exploitation. With this set-up, many believe that the ability of blockchain to provide total control will help increase feelings of safety, and, hence, trust.
Isn’t that what managers should be doing in the first place? If so, technology like blockchain will indeed become part of our management systems very soon.
Management by algorithm
But, to answer the question of whether the algorithmic manager will wake up soon, let us return again to how we defined management. As I explained earlier, the purpose of management is to ensure that order and stability is maintained. According to our initial analysis, algorithms seem perfectly equipped to achieve this purpose. Indeed, as all the examples illustrate, algorithms penetrate managerial jobs by providing more specific and consistent ways of assessing, monitoring and evaluating employees.
Today, algorithms are able to learn about employees in fast and accurate ways, and can make valuable and reliable predictions regarding the future behaviors of those employees. As such, algorithms do possess the necessary skills to execute many – if not most – managerial tasks. It is at this stage that we can say that management by algorithm has become a reality and will only grow further in importance in managing our future workforce. Management by algorithm is not a fantasy anymore, it has arrived and is likely to stay.
The next steps in allowing algorithms to penetrate management functions are already being prepared. At first, we saw algorithms capable of executing monitoring functions at large, but today, we are already moving into a more complex reality where algorithms can take over actual managerial decision-making roles. Along with an increasing technological sophistication, which allows algorithms to learn on their own (without human involvement and supervision), algorithms are gradually replacing human resources to increase efficiency of execution and productivity. One example is that algorithms are moving more gradually into advisor roles. In this role they provide feedback on how human employees have to interpret data analytics and what those interpretations mean in light of the decisions that have to be taken. Take for example, the situation where algorithms are being prepared to analyze the skill sets employees need to perform in the organization of tomorrow, but at the same time also make suggestions on the appropriate pay levels for the employees that they evaluate.
At this point, you may say: hold on, the influence of algorithms may technically be possible, but will we simply accept their advice? As I mentioned earlier, people can be OK with algorithms making decisions. In some cases, they may even prefer it, because algorithms are not seen to have their own intentions and therefore their decisions are considered unbiased.
So, overall, the involvement of algorithms in managerial decision-making clearly brings benefits to the table and, interestingly, we may well be ready to accept this. At the same time, however, we have to remain aware of where the limits lie in accepting algorithms as a decision-making tool. But, why should we be aware of those limits? Isn’t it a good thing that algorithms bring more objectivity into our decision-making and, as such, relieve humans from making tricky decisions where the risk of damaging the interests of others is high?
True, increasing the objectivity of any decision is a gain. But, by placing algorithms in a role where they are the ones deciding on the interpretation of the data, the risk exists that once we do not understand anymore how the advice comes about, we are not simply guided by algorithms, but led by them. Indeed, if we reach a point where algorithms are self-learning, to the extent that we do not understand any more why certain decisions have to be made, we will have given the responsibility of decision-making to machines.
Being responsible for a decision also means that we need to make sure that the interests of all stakeholders are served. In other words, the responsibility of decision making lies in being able to make the right judgment calls to ensure any value created contributes to the welfare of others. If algorithms can grow into this role, we need to be sure that they will make human-centered choices which respect the welfare and protect the interest of humans. After all, if algorithms reach the point of self-learning and decision making, we have little or even no latitude to correct.
The limits of management by algorithm
Because of these reasons, I feel that we need to think not only about the unlimited potential for algorithms to make decisions, but also about where we set the limits of an algorithm’s decision-making authority.
As I will discuss later, algorithms are limited in taking the perspective of a human and understanding the deeper emotions that underlie the human ability to make judgment calls. But, aside from the fact that algorithms do not have a real understanding of what the human condition entails, they are also vulnerable to certain errors and biases that could violate respect for the human identity. As a result, algorithms may possess the ability to optimize decision making, but a lack of empathy and understanding of what it means to be human will increase the risk of omission errors, in which alternatives that are more human-centered may be discarded.
For these reasons, we need to decide on how best to regulate self-learning technology, which implies deciding on where the limits of management by algorithm lie. Although humans also display biases, we are usually aware of them and have the ability to empathize with those being treated unfairly. Hence, a human would likely try to correct the situation. As algorithms do not experience such compassion, their more accurate, even superior, decision-making abilities will not allow them to identify biased implications towards human actors.
Indeed, growing evidence suggests that algorithm-based decisions amplify human biases in the data they analyze. The problem is that while humans recognize a bias, algorithms do not, because they learn from observation and observable trends. The meaning and emotional connotation behind such trends will not be recognized by an algorithm and thus, decisions will not be assessed and revoked when necessary. For example, in 2018, Amazon applied algorithms in their hiring decisions and discovered that the algorithm gave higher scores to white males.
So, the algorithm was making decisions in favour of white humans who were male. How is this possible? Well, the algorithm learned from historical job-performance data and recognized the trend that in the past white men had been the best performers. Humans obviously know that in the past the majority of those employed were white men, but that norms regarding diversity in society has now changed. It is humans who have that emotional awareness, but algorithms do not. Hence, it does not make them able to make judgment calls. Of course, as I mentioned earlier, humans are also capable of making biased decision, but because of the human ability to create meaning from different perspectives, remedies and solutions can be identified. It is therefore no real surprise that as soon as Amazon identified this problem, the company stopped using the system since there was no simple way to fix it.
So, where do the limits lie of management by algorithm? The first conclusion is that many, if not most, managerial tasks that fit our definition of management will become automated. Management represents a way of providing a stable work environment and algorithms can execute most of these tasks. Does this mean that we can then simply delegate the responsibility of managerial tasks and decisions fully to algorithms? No, it does not. What we can learn from Amazon’s failure is that execution of management can be done by algorithms, but not without human involvement and oversight.
Increased speed, heightened accuracy and replicability of decision outcomes is a very valuable addition to the management process, and algorithms are the ideal candidates to perform such tasks. But, when it comes down to making decisions where different human-stakeholders are involved, companies need to realize that the interpretability of the decision-making process is more important than the skills algorithms bring with them. Indeed, decisions need to be put in context and interpreted in light of different perspectives. As a result, algorithms cannot move into autonomous management roles where they assume full responsibility for decisions made.
Co-operation is the key word
Management by algorithm is a reality, but it is constrained to the automation of managerial execution efforts. In addition, managerial tasks and decision making that does not impact the value of the company may, over time, also be delegated to algorithms. All of this also makes sense in light of how companies want to position themselves in the market.
Think about it. If we see automation of management becoming possible, and in this process we decide to delegate increasingly more responsibility and authority to algorithms, then all of us will eventually adopt similar approaches to the data out there. As a result, we will increasingly make similar decisions, which will lead to the reality that companies become almost replications of themselves. Running companies in unlimited automated ways will therefore increase the likelihood that all those companies will become interchangeable! However, is business not about being able to create value based on unique priorities and strategies that distinguish you from the competition? Yes, it is.
If we are afraid that automation of our management will make us dependent on what the algorithm delivers, then we put ourselves in a position where we look at companies as passive recipients of automated authorities, spitting out data-driven advice. Unfortunately, no passive company can be creative, forward looking, and create value for society and the stakeholders it serves. The reality is, if our organizations were to be run in this way, then it is likely that we may destroy many beautiful assets and resources (e.g. a collaborative, trusting and empowering work context), compared to organizations led by leaders with clear values on how to deal with different stakeholders.
Some of my own recent research points out this possible failure.104 In a series of experiments, we placed human participants in a work context where their task performance was assessed and evaluated by an autonomous algorithm. The evaluation of the algorithm was used directly by the top of the company to reward the participant. No intervention of a human supervisor was possible. No possibility was given to participants to express one’s view or feelings towards human supervisors. We also created another condition where an autonomous algorithm performed the same assessment, but this time our participants could talk to a human supervisor and share their experiences. What we found was that in such an automated performance context, human participants considered the work context to be more trustworthy and fair when human supervisors were involved. This was particularly the case when those supervisors were also perceived as being humble (and thus value-driven).
What these results indicate is that, in a context where algorithms engage in the managerial task of conducting performance evaluations, employees feel that human leadership with strong values is needed to create a sense of fairness and trust in how the organization as a whole operates.
Although the reality of management by algorithm makes it clear that management functions are likely to be replicated by algorithms, it also makes clear where the future of human management lies. That future is in the domain of leadership. It appears that when automation goes up, so do a number of other factors: our need to have leadership in place that knows what it wants to achieve; leadership that offers judgment when decisions have to be taken; and leadership that can reflect effectively on the goals to be pursued.
Data is all well and good, and can point out several directions to take, but the final, strategic decision lies with human leadership. Because of the unique human capabilities, the direction a company takes ultimately always needs to be based on what is valuable to the business, their customers and society. The reality of running an organization in the future will be that the more algorithms take over management, the more we will need leadership to bring in human judgment to help set priorities. If this is the case, then we see a similar story emerge, as we did in the past, where we emphasized that the running of an organization requires managers and leaders to work together to create both innovative and sustainable ways of doing business.
In today’s business climate, we again see this pattern emerge through algorithms taking care of management and humans accounting for the leadership responsibilities. This idea indicates that the way to run organizations in the future will be to follow a collaboration model, where algorithms and humans jointly create value. As Kevin Kelly in his book, The Inevitable, notes: “This is not a race against the machines … This is a race with the machines.”105
89 Tarnoff, B. (2017). ‘Silicon Valley siphons our data like oil. But the deepest drilling has just begun.’ The Guardian. Retrieved from: https://www.theguardian.com/world/2017/aug/23/silicon-valley-big-data-extraction-amazon-whole-foods-facebook
90 Thorp, J. (2012). ‘Big data is not the new oil.’ Harvard Business Review. November 30. Retrieved from: https://hbr.org/2012/11/data-humans-and-the-new-oil
91 Lapuschkin, S., Wäldchen, S., Binder, A., Montavon, G., Samek, W., & Mueller, K.-R. (2019). ‘Unmasking clever Hans predictors and assessing what machines really learn.’ Nature Communications, 10, 1096.
92 Editorial (2017). ‘How does the brain work?’ Neuron, 94(5), 933.
93 Goldhill, O., (2015). ‘Algorithms make better hiring decisions than humans.’ Retrieved from: https://qz.com/561206/algorithms-make-better-hiring-decisions-than-humans/
94 Fly, A. (2019). ‘The skills leaders need to survive in the age of AI,’ https://www.techradar.com/sg/news/the-skills-leaders-need-to-survive-in-the-age-of-ai)
95 IBM (2018). ‘Power your candidate experience with AI.’ Retrieved from: https://newsroom.ibm.com/IBM-watson?item=30401 (also, Leight-Deobald et al., (in press). ‘The challenges of algorithm-based HR decision-making for personal integrity.’ Journal of Business Ethics.)
96 Bonczek, R.H., Holsapple, C.W., & Whinston, A.B. (1979). ‘Computer-based support of organizational decision making.’ Decision Sciences, 10(2), 268-291.
97 Courtney, J.F. (2001). ‘Decision making and knowledge management in inquiring organizations: Toward a new decision-making paradigm for DSS.’ Decision Support Systems, 31(1), 17-38.
98 Gigerenzer, G., & Gaissmaier, W. (2011). ‘Heuristic Decision Making.’ Annual Review of Psychology, 62, 451-482.
99 Huber, G. (1990). ‘A theory of the effects of advanced information technologies on organizational design, intelligence and decision making,’ Academy of Management Review, 15(1), 47-71.
100 Pomerol, J.-C. (1997). ‘Artificial intelligence and human decision making.’ European Journal of Operational Research, 99(1), 3-25.
101 Russel, S.J., & Norvig, P. (2016). ‘Artificial intelligence: A modern approach.’ Pearson Education Limited.
102 Rosenblat, A., Kneese, T., & Boyd, D. (2014). ‘Workplace surveillance.’ Data & Society Working Paper. New York: Data & Society Research Institute.
103 Volini, E., Schwartz, J., Roy, I., Hauptmann, M., Van Durme, Y., Denny, B., & Bersin, J. (2019). ‘Organizational performance: It’s a team sport.’ Deloitte report, 2019 Global Human Capital Trends. April 11. Retrieved from: https://trendsapp.deloitte.com/reports/2019/global-human-capital-trends/organizational-performance.html
104 De Cremer, D., McGuire, J., Naryanan, J., Tang, P., & Mai, M. K. (2020). ‘How fair and trustworthy automated assessment systems are: it depends on you speaking up and your supervisor’s humility.’ Working paper NUS Business School.
105 Kelly, K. (2016). ‘The inevitable: Understanding the 12 technological forces that will shape our future.’ Viking Press.