Human beings are the ghosts in the machines—independent forces in all the automata, machines and computers ever created. If that were not the case, could the robot apocalypse foretold in the 2004 cult classic film I, Robot be a premonition of a future now only 13 years away?1
Either way, machines continue to take over more jobs at the expense of people. The situation is exacerbated by the way enterprises see the value in people and how traditional enterprises maintain structures in the context of the accelerated rise of machines in the 21st century.
Digitally native enterprises have found that there are better ways to organize their people than the haunted and dated organizational design paradigms that were first practiced long before the rise of the modern computing driving today’s machines.
Currently, all the machines that exist, digital or otherwise, do their humans’ bidding while suffering from many of the same things that limit human beings, such as bias.
Background
Long before “ghost in the machine” became an English idiom, it was a derogatory term used by British philosopher Gilbert Ryle to describe the work of French philosopher and mathematician René Descartes.2 Descartes was more than just interested in the machines known as automata; he designed and built them.3
As he explored the relationship between mind and body, Descartes proposed that the mind is not physical; it exists independently of the brain, with the body essentially being a biological automaton. The concept that machines lack something humans have—a mind—is the basis for the Turing test, the traditional assessment of whether a computer can think and act in a way indistinguishable from a human.4 However, Ryle called Descartes’ mind-body dualism “the ghost in the machine,” arguing that the mind is dependent on the brain.5
Whether the mind exists independently of the body is not at issue here; rather, the question is the nature of the relationship between humans and machines. In 1863, British author Samuel Butler wrote that machines would one day overtake humans as the superior race6 at a point referred to today as the technological singularity.7 Currently, all the machines that exist, digital or otherwise, do their humans’ bidding while suffering from many of the same things that limit human beings, such as bias.
The shift to post-industrial economics is driving the need for new thinking about the organizational design requirements of post-industrial enterprises.
The Machines
When talking about technology, machines or automata, it is forgivable to believe that technology does much more for humankind than it really does. Smart contracts in blockchain technology handle supply chain contracts; software enacts IT governance, risk management, and compliance (GRC) policies; robots driven by artificial intelligence (AI) seem to have intelligent transactional conversations with bank customers; and machines man the checkout lines at leading retailers at the points of sale (POS). It seems there is little left for humans to do. Even self-driving cars have automated the operation of steering wheels, accelerators and brake pedals and all the sensory inputs required to get from point A to point B.
If English mathematicians Charles Babbage and Ada Lovelace were alive today, they would argue that computer-based automation is nothing new. The technology is 200 years old, thanks to Babbage’s analytical engine (computer) and Lovelace’s punch cards (program) that made the engine compute (figure 1). So, a recent claim that automation will be a “global force that will transform economies and the workforce”8 is an interesting view of the future given that this transformation has already long been in force.
Source: Based on Institute of Chartered Accountants in England and Wales, (ICAEW), “History of Blockchain,” http://www.icaew.com/technical/technology/blockchain-and-cryptoassets/blockchain-articles/what-is-blockchain/history#:~:text=Blockchain%20has%20the%20potential%20to,the%20pseudonym%20of%20Satoshi%20Nakamoto; Foote, K. D.; “A Brief History of the Internet of Things,” Dataversity, 14 January 2022, http://www.dataversity.net/brief-history-internet-things/; Perry, D. G.; S. H. Blumenthal; R. M. Hinden; “The ARPANET and the DARPA Internet,” Library Hi Tech, vol. 6, iss. 2, 1 February 1988, http://www.emerald.com/insight/content/doi/10.1108/eb047726/full/html?skipTracking=true; Anyoha, R.; “The History of Artificial Intelligence” Harvard University, The Graduate School of Arts and Sciences, Cambridge, Massachusetts, USA, 28 August 2018, http://sitn.hms.harvard.edu/flash/2017/history-artificial-intelligence/; Stanford Computer Science, “Robotics: A Brief History,” Stanford University, California, USA, http://cs.stanford.edu/people/eroberts/courses/soco/projects/1998-99/robotics/history.html#:~:text=The%20earliest%20robots%20as%20we,industry%2C%20but%20did%20not%20succeed; The Editors of Encyclopedia Britannica; “Personal Computer,” Encyclopedia Britannica, http://www.britannica.com/technology/personal-computer; Staff, “Generations, Computers” Encyclopedia.com, http://www.encyclopedia.com/computing/news-wires-white-papers-and-books/generations-computers; The Editors of Encyclopedia Britannica; “Charles Babbage,” Encyclopedia Britannica, http://www.britannica.com/biography/Charles-Babbage; History.com Editors; “Morse Code and the Telegraph,” History.com, 6 June 2019, http://www.history.com/topics/inventions/telegraph#:~:text=Developed%20in%20the%201830s%20and,a%20wire%20laid%20between%20stations
For example, automation has already transformed economies, moving “us from a craft system to mass production, from blue-collar to white-collar to ‘new-collar’ work—with better work, higher wages, more jobs, and better living standards.”9 In addition, automation has already been transforming workforces for centuries, with examples occurring in the 16th century (stockings), 19th century (textiles) and 20th century (automobile manufacturing).10 The textile automation of the 19th century by means of the Jacquard loom (which used punch cards to automate the woven patterns) is what inspired Babbage’s analytical engine, and Lovelace’s punch-card mechanism was still programming computers 150 years later.
The shift to post-industrial economics is driving the need for new thinking about the organizational design requirements of post-industrial enterprises.
As shown in figure 1, robots and AI have been present since the 1950s and continue to have an impact on the nature of work. The public Internet changed the nature of office memoranda and the postal service and grew into a much larger capability than ever imagined, with both good and bad results. The Internet of Things (IoT) (e.g., sensor technology) came on the scene in the late 20th century, with recent developments occurring in the field of autonomous supply chain logistics, for example. The rise of blockchain in the early 21st century has already made an impact, such as by automating compliance with legal contracts. Computing machines will continue to lead the automation revolution, with AI and robotics evolving to better imitate humans, and automation will continue to transform economies and workforces.
Today, IT is still about automation. Smart contracts are the automated version of manual contract validation. GRC software automates steps to meet IT policy requirements that would otherwise be executed manually. The robot automates the bank teller’s activities, a machine automates the process of checking out at the retail store, and self-driving cars automate driving. These are just some of the machines in use today.
The Ghosts
Human beings are still in charge of deciding what will be automated and how automation will be implemented. Humans develop the smart contracts, and humans define how they will work. Humans write the GRC software, and humans decide which events that software will address. Humans program the robot, and humans decide in what languages the robot can communicate. Humans install and maintain the retail POS machine, and humans perform the simple task of refilling it with paper to dispense receipts. Humans manufacture the sensory devices required for self-driving cars, and humans develop the routines to check them.
With the introduction of accessible and more user-friendly personal computer (PC) technology (vs. mainframes), task automation began to increase.
One of the important differences between humans and machines is that humans have the ability to understand various situations and behave accordingly, while machines do not.11 In other words, humans can respond to change, whereas machines cannot. In closed loop systems, machines can react to a limited set of changes because of the feedback loop in the cycle, but in general, machines need humans to change their configurations so they can operate appropriately in environments that exceed the boundaries of the system’s original specifications.
Some may argue that machines based on AI—or, more specifically, machine learning (ML) as a subset of AI—can respond to changes in the model’s underlying data. However, changes in the underlying data resulting from a change in their operating environment and models result in model drift. That is, a model may no longer adequately represent the environment for which it was originally designed and tested. So, not even ML machines can change and test themselves in all circumstances without human intervention.
Humans are still the driving force in even the most modern enterprises. Without humans, machines would be unable to adjust to changing circumstances and would, therefore, lose their value to society. Without machines, though, humans could adjust and survive, as they once did. Humans are still the ghosts in the machines.
Working Ghosts in an Era of Machines
The history of work can be summarized as humankind’s journey from hunter-gatherers and nomads to vocations resulting from the agricultural revolution that started 12,000 years ago to the blue collar labor of 18th-century industry and to the white collar service industries of the 20th and 21st centuries.12 Reflecting the trend in work shifting from blue collar to more white and new collar, about a third of the Canadian working population was employed in the service sector in 1911, and by 1987, about two thirds were employed in that sector,13 showing the extent of the migration. This example highlights the need for new organizational design thinking fit for a post-industrial economy.
Rapid technological developments over the last decade alone have demanded significant increases in organizational agility.
Computerization entered the workplace in the 1960s with second-generation computers (those using transistors instead of the valves of first-generation computers), focusing on “accounting, and in the processing of payroll, personnel, and inventory records.”14 With the introduction of accessible and more user-friendly personal computer (PC) technology (vs. mainframes), task automation began to increase. The impact of computing technology included a reduction in low- and mid-level clerical work and the introduction of new types of work, given the increasing availability of information at all organizational levels.15
Computing technologies have had major impacts on workflow, job content and job responsibilities. Since the 1960s, for example, organizational structures have been influenced by computer centers that facilitated computation and reporting.16 If enterprises were already undergoing change 60 years ago, how have new generations of emerging technology, including AI, which has increased in maturity since the 1950s, affected organizational design in the 21st century?
In terms of the machines themselves, completely roboticized and automated production is not yet a reality. The first generation of industrial robots could reliably perform simple tasks only if their environments were identical for each iteration of the task. Although computer scientists and engineers have developed more sensitive robots to manage environmental variances, humans are still needed to support and maintain the robots’ productivity, another clear and current reminder that humans are the ghosts in the machines.17
How can humans group and organize themselves in the presence of post-industrial machines to better perform their work, some of it novel with respect to any work that might have come before?
The Problem of Old Ghosts Haunting New Ghosts
Organizational functional design (e.g., a finance department) and divisional design (e.g., a personal banking division in a large bank) are relics of 1908 and 1925, respectively, but many of today’s hottest topics in organizational design are even older (figure 2).18
Adapted from Denison, D. R.; A. Narasimhan; M. J. Piskorski; “Designing a High Performance Organization,” IMD, March 2015, http://www.imd.org/research-knowledge/articles/designing-a-high-performance-organization/
Organizational culture—the norms and behaviors experienced in the workplace every day—originated in the year 696.19 The management of knowledge and intellectual property was an organizational characteristic in 1149, while project-based teams have been topical since the 16th century and matrix organizations have existed since 196920 (figure 2). Somewhat more modern are self-managed teams and customer-centricity, which have been around since 1971 and 1992, respectively.21 However, rapid technological developments over the last decade alone have demanded significant increases in organizational agility. Thus, are any of these structures—the youngest being 30 years old—really up to the task of handling the scaling, diversification and responsiveness required to manage the work of people and machines in the 21st century?
Answering this question would require an analysis of humans in society over time. The societal influence on work means that it can be defined as the exploitation of one class over another, as achievement or even as punishment.22 Functionally, double-entry accounting and the first modern enterprises were present during the Renaissance (16th and 17th centuries), while the Industrial Revolution (18th and 19th centuries) brought the division of labor and the distinction between white-collar (intellectual) and blue-collar (manual) work, terms still in use today.23 The job was born in the 19th century, as was bureaucracy, introduced by German sociologist Max Weber, who described the type of structured, formalized, impersonal and hierarchical organizations that exist today (figure 3).24
US mechanical engineer Frederick Taylor introduced scientific management, with a focus on human efficiency and productivity. Automation has been a significant factor in achieving organizational efficiency and productivity since then. The work of Weber and Taylor still survive 100 years later, while the most modern organizational structure, customer-centricity, is 30 years old.
The method of organizing people for work is dated, based on thinking by people who have long since crossed the threshold into the spirit (ghost) world.
It is not hard to see the problem: How can even the most modern general organizational design pattern or management paradigm fulfill 21st-century human organizational requirements and keep up with the exponential rate of computing machine development when it is already so much older than the reigning technologies of the day? The definition of the job and of organizational bureaucracy entered the scene before computer automation and the development of the tertiary and quaternary sectors (figure 3).
What has changed since? Customer-centricity as an organizational paradigm is new in the age of growing white collar and new-collar work for the services and knowledge sectors respectively.
Enterprises are still largely structured like the physical industrial factories of 100 years ago—with divisions and functions—perhaps implicitly assuming that these are still the best structures for people increasingly employed in the service sector in an increasingly digital world. The method of organizing people for work is dated, based on thinking by people who have long since crossed the threshold into the spirit (ghost) world—a case of old ghosts haunting the new ghosts in the machines.
Exorcising the Old Ghosts
Work arrangements such as offshoring, outsourcing, working from home and crowdsourcing (all relatively recent constructs) require new skills, which has consequences for the structure of work and employment.25 None of these factors were in place to influence the design of today’s dated organizational structure paradigms. The same is true for automation. In examining the modern working relationships between people and robots—such as in automotive assembly—people constantly need to acquire new qualifications and capabilities because system integrators and robot manufacturers seldom consider the social impact of new technologies,26 never mind the organizational impact.
That humans are changing more slowly than the pace of automation evolution provides a modern context for thinking about Butler’s 150-year-old singularity:27 Does the slow rate of human change, coupled with exponential increases in technological capability, mean that the hypothetical technological singularity is closer than people think?
Other work- and structure-related issues arising from automation in the 21st century include:28
- Less-skilled and lower-educated workers are losing their jobs to automation. This continues to feed inequalities in the workplace, particularly for women, the young, the old, and those without post-school qualifications.
- As technology advances, even previously nonautomatable jobs held by low-skilled, low-educated workers are being performed by machines. This means that new technologies are increasingly skill biased rather than routine biased, putting the jobs of already disadvantaged workers at risk of disappearing.
- There has been a significant increase in repetitiveness and standardization in the work of highly skilled workers, suggesting that automation is commoditizing even previously nonroutine cognitive tasks.
Enterprises structured by the old ghosts need to digitally transform to maintain their market relevance, in contrast to just being digital—the hallmark of so-called digital-native enterprises. Digital-native enterprises started out as digital; they did not need to go through digital transformation to become digital.
A characteristic of digital-native enterprises is that the motivation to:
[C]ontinuously experiment, learn and adapt [has] been embedded in their organizational DNA. They have built a continuous cycle of digital innovation…[to] capture existing markets and create new ones, and continue to attract talent and investment.29
A key factor is that digital-native enterprises are “customer- and product-led, using enterprise technology as an enabler and empowering all employees to innovate by design.”30
Many factors differentiate digital-native enterprises from traditional ones (figure 4).31 For traditional enterprises, successful digital transformation requires due consideration of capacity, capability and structure with respect to performing work digitally. Beyond organizational capacity and capability, however, the organizational design and structure must ensure that digital work is appropriately divided throughout the enterprise, with managed duplication, clear differentiation of accountabilities and responsibilities, and flexibility.32 The challenge for many of today’s organizations is to evolve their old organizational structures—represented in the left column in figure 4—by integrating the organizational demands of the modern organization—represented in the right column of figure 4.
Conclusion
Traditional enterprises’ focus on skill sets vs. mindsets is not a positive sign for their staffs. This is because skilled labor is now the target of automation, and large international management consulting enterprises continue to push automation in search of greater productivity and efficiency.
Ensuring that people in skilled positions remain employable over time means not only increasing their skills, but also encouraging mindsets of innovation— for example, customer-centricity and comfort with managed failure—organizational cultural constructs that are currently difficult for machines to replicate. Focusing on skills growth alone may be a losing battle against the machines and, potentially, against the singularity.
Organizational culture is a 1,300-year-old paradigm. Digitally native enterprises are not only flexible in their constitution, but also connect to other teams as and when needed and without an explicit hierarchy. The 16th-century construct of team-based structures is quite possibly the greatest visible differentiator between traditional and modern enterprises. The implications of staid structures and management paradigms for the sustainability of traditional enterprises are significant.
Enterprises decide to pursue digital transformation for many reasons. However, traditional enterprises are in a catch-22 situation: Do not change, and risk becoming irrelevant; change, and be subject to significantly higher risk related to the need to navigate interconnected legacy systems, legacy processes and legacy thinking. The implication of this finding is a critical need for active risk management for established, traditional enterprises pursuing digital transformation.
Ensuring that people in skilled positions remain employable over time means not only increasing their skills, but also encouraging mindsets of innovation.
That digital-native enterprises focus on mindsets relative to the digital machines they create is reminiscent of the difference between Descartes’ mind and automata. Perhaps Descartes was not far off the mark, and Ryle’s derogatory reference to “the ghost in the machine” was really a toast to the power and sustainability of Descartes’ thinking 400 years ago. Ultimately, humans can be more effective if they are not held back by old, haunted organizational structures compromising their growth and development. The old ghosts can be exorcised if care is taken, and humans can maintain their position as ghosts in the machines, at least for the foreseeable future.
Endnotes
1 Proyas, Alex; I, Robot, 20th Century Studios, Los Angeles, California, USA, 2004
2 Ather, S. H.; “A History of Artificial Intelligence,” http://ahistoryofai.com/descartes-2/
3 Ibid.
4 Ibid.
5 Grammarist, “Ghost in the Machine,” http://grammarist.com/idiom/ghost-in-the-machine/
6 Danaylov, N.; “17 Definitions of the Technological Singularity,” Institute for Ethics and Emerging Technologies, 22 August 2012, http://archive.ieet.org/articles/danaylov20120822.html
7 Kurzwell, R.; “Singularity: Explain It to Me Like I’m Five-Years-Old,” Futurism, 3 March 2017, http://futurism.com/singularity-explain-it-to-me-like-im-5-years-old
8 Edlich, A.; G. Phalin; R. Jogani; S. Kaniyar; Driving Impact at Scale From Automation and AI, McKinsey, February 2019, http://www.mckinsey.com/~/media/McKinsey/Business%20Functions/McKinsey%20Digital/Our%20Insights/Driving%20impact%20at%20scale%20from%20automation%20and%20AI/Driving-impact-at-scale-from-automation-and-AI.ashx
9 Fitzpayne, A.; C. McKay; E. Pollack; Automation and a Changing Economy: The Case for Action, Future of Work Initiative, Aspen Institute, USA, 2 April 2019, http://www.aspeninstitute.org/publications/automation-and-a-changing-economy-the-case-for-action/#:~:text=Automation%20helped%20move%20us%20from,jobs%2C%20and%20better%20living%20standards
10 Fleming, S.; “A Short History of Jobs and Automation,” World Economic Forum, 3 September 2020, http://www.weforum.org/agenda/2020/09/short-history-jobs-automation/
11 Prabhat, S.; “Difference Between Human and Machine,” Difference Between Similar Terms and Objects, http://www.differencebetween.net/miscellaneous/difference-between-human-and-machine/#ixzz7Reph9y7K
12 Holmer, A.; “A History of Work: Livelihood, Vocation, and Labour,” WorkMatters, 10 April 2020, http://medium.com/workmatters/the-history-of-work-livelihood-vocation-and-labour-32196fe41c54
13 Phillips, R. A.; “Service Industry,” The Canadian Encyclopedia, 4 March 2015, http://www.thecanadianencyclopedia.ca/en/article/service-industry
14 Stanback, T. M., Jr.; Computerization and the Transformation of Employment, Routledge, USA, 1987, http://www.routledge.com/Computerization-And-The-Transformation-Of-Employment-Government-Hospitals/Stanback/p/book/9780367163570
15 Ibid.
16 Ibid.
17 Kransberg, M.; M. T. Hannan; “Automation,” Britannica, 1 November 2021, http://www.britannica.com/topic/history-of-work-organization-648000/additional-info#contributors
18 Denison, D. R.; A. Narasimhan; M. J. Piskorski; “Designing a High Performance Organization,” IMD, March 2015, http://www.imd.org/research-knowledge/articles/designing-a-high-performance-organization/
19 Ibid.
20 Ibid.
21 Ibid.
22 Caredda, S.; “Part 1: A Brief History of Work,” Sergiocaredda.eu, 18 October 2020, http://sergiocaredda.eu/people/future-of-work/part-1-a-brief-history-of-work/
23 Ibid.
24 Ibid.
25 Ozkiziltan, D.; A. Hassel; “Humans Versus Machines: An Overview of Research on the Effects of Automation of Work,” SSRN, 8 August 2020, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=3789992
26 Ibid.
27 Ibid.
28 Ibid.
29 Watt, M.; “How Digital Natives Are Influencing Traditional Organizational Design,” EY, 22 July 2020, http://www.ey.com/en_gl/workforce/how-digital-natives-are-influencing-traditional-organizational-design
30 Ibid.
31 Ibid.
32 Pearce, G.; “Attaining Digital Transformation Readiness,” ISACA® Journal, vol. 1, 2020, http://wup.ozone-1.com/archives
GUY PEARCE | CGEIT, CDPSE
Has an academic background in computer science and commerce and has served in strategic leadership, IT governance, and enterprise governance capacities, mainly in financial services. He has been active in digital transformation since 1999, focusing on the people and process integration of emerging technology into organizations to ensure effective adoption. Pearce was first exposed to artificial intelligence (AI) in 1989, and he has followed the evolution of the discipline from symbolic AI to statistical AI during the intervening decades. He was awarded the 2019 ISACA® Michael Cangemi Best Author award for contributions to IT governance, and he consults in digital transformation, data governance and IT governance.