![Sample our Social Sciences journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110822/TOC-Subject-Sample-2021-Social-Sciences.jpg"})
Abstract
The fourth industrial revolution has become a prominent concept and imminent technological change a major issue. Facets are everyone’s concern but currently no one’s ultimate responsibility (perhaps a little like financial stability before the global financial crisis). In this paper, we argue that the future is being shaped now by the way the fourth industrial revolution is being positioned. Whilst no one has set out to argue for or defend technological determinism, anxiety combined with passivity and complacency are being produced in the context of a quasi-determinism. The contingent quantification of the future with regard to the potential for job displacement provides an influential source of authority for this position. A background narrative of ‘the future is coming, so you better get used to it’ is being disseminated. This favours a capitalism that may ‘deny work to the many’ perspective rather than a more fundamental rethink that encompasses change that may liberate the many from work. This, in turn, positions workers and responsibility for future employment, reducing the urgency of calls for wider societal preparation. Public understanding and policy are thus affected and along with them the future of work.
Acknowledgements
Thanks to Steve Fleetwood for support with this paper and Andrew Brown and Bob Jessop for valuable comments on an initial extended version. Thanks also to the anonymous reviewers.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1 There are various issues to disentangle. For example, globalization theory is underpinned by comparative advantage, but the theory takes an implicitly consumer point of view, and tends to ignore distributions and real time effects, whilst assuming capital is fungible rather than specific, static and destroyed if corporations relocate. As ideology, this conflates protectionism with the creation of domestic policy space to shape trade, and enhances both the power of MNEs and the scope of financialisation (see Fullbrook & Morgan, Citation2017).
2 The issues are complex. For example, a focus on bringing factories back needs to carefully specify that assembly is a small fraction of the overall production process. Apple, for example, uses over 750 contracted entities around the world to produce an iPhone and only $8 of the estimated $378 cost of an iPhone X is attributable to final assembly. Moreover, this aspect of the process is most easily automated and least likely a source of secure high wage employment. Foxconn in China is already heavily investing in automation so the gains from any relocation into Europe (the United Kingdom) or United States might be small and, if one were to justify this in terms of redressing balance of trade issues, this seems more a matter of the failure of standard measures of trade to be calculated in terms of the value added in places rather than the total valuation of products shipped (the US deficit with China, for example, is likely a third lower than the recent figures that have exercised President Trump).
3 AI discourse learning, intelligence, etc. have been appropriated as reasonable terms to use regarding what machines/computers/programmes do. This, of course, disguises a basic difference to the human who has intelligence or can learn - consciousness and self-consciousness, which makes the entity not just a system of symbol manipulation for functions but a being for whom processes are meaningful. This has created significant debate in the philosophy of AI initiated by Turing and Searle (Morgan, Citation2018b). ML and AI research has moved on from simple discrete-state input-output concepts and approaches, and Bayesian or Boolean solutions. The major innovation providing the background in current AI is ‘deep learning’ using artificial neural networks (ANN). ANN are described as software simulations of neuron connectivity (The Economist, Citation2016). That is, they are multiply layered sets of ‘neural units’ creating multiple dividing points for direction, as processing, from some given input to some output. The sophistication of the system or its capacity for difference and range is based on the number of layers, the ‘depth’, in the structure. What the system is directed to can then (currently) be set up in three ways expressed as learning modes: (1) supervised learning (a network system is fed an example dataset that exemplifies what it is intended to achieve, such as spam identification) (2) unsupervised learning (a network system is fed an example dataset and is set up to look for patterns, clusters, anomalies in the data, which then become the specific output within a broader data-defined remit, such as fraud patterns in insurance claims) (3) reinforcement learning (a network system is fed an example dataset and refines its behaviour based on rewards as feedback to achieve goals, creating a simulation of ‘do what works best in situation x’, such as playing and winning a video game). In all three cases the key innovation is that the network progressively refines the weighting between connections, and it thus fine-tunes the network system. The more data the system has to work with, the more layers to the neural network and the more simulations run, then the more effective the system becomes, over time and in real time, subject to processing capacity and speed.
4 For example, the multinational enterprise ABB dominates the production and development of industrial robotics (its smallest of four divisions is larger in revenue terms than the next four largest corporations combined). ABB’s Yumi range next generation robot is networked, sensor-fitted, dual-armed, multi-functional, easily reprogrammable, only 38 kg and cost $40,000 per unit. Software support is also provided for hardware systems to allow virtual factory redesigns to improve (essentially Taylorist) production systems and to enable anticipation of mechanical problems based on analytics of wear and tear, etc. The implication is that future factories can be small and flexible and require far lower initial investment. All the main corporations are developing similar ranges and support services.
See http://new.abb.com and http://new.abb.com/future ABB is one of the listed ‘partners’ of WEF’s Centre for the Fourth Industrial Revolution, https://www.weforum.org/centre-for-the-fourth-industrial-revolution/about
5 This in turn is reflected in the primary research published across the sciences, which intersects with operational research and technocratic organizational modelling problem sets, and this is where much of the formal work is currently being done making use of the fourth industrial revolution concept (typically as industry 4.0). See recent work in the journals: Cybernetics and Systems, International Journal of Computer Integrated Manufacturing and Production Planning and Control. The range of current projects and programmes is set out most clearly in the World Economic Forum Deep shift report (WEF, Citation2015). The associated forecasts based on surveys are, however, highly contestable.
6 For example, Boston Dynamics Youtube video is distributed via The Guardian newspaper’s site: https://www.theguardian.com/technology/video/2018/feb/21/human-robot-dog-boston-dynamics-door-opening-spotmini
7 See also the WEF ‘Fourth industrial revolution for the Earth’ series: https://www.weforum.org/agenda/2018/09/can-technology-save-life-on-earth
8 This is to say nothing of the inherent dangers of a new phase in dependency: a society of division of labour creates a situation where we become mutually dependent and begin to lose the skills and capacity to survive in small groups or isolation, an increasing use of technologies and then delegation of activity to technological systems exacerbates the basic problem of how one survives any significant dysfunction to the system we call civilization. This is slightly different than AI singularity and Terminator scenarios.
9 For example, according to the International Federation of Robotics there are currently fewer than 2 million industrial robots in the world and the vast majority are purposed for the automotive industry. Current investment trends indicate the growth in use of industrial robots is highly variable by industrial sector and geographical region (China and East Asia dominate based on production, investment and density). However, the eventual impact on jobs is considered to be significant and draws heavily on research findings from the main fourth industrial revolution literature. See IFR (Citation2017a, Citation2017b).
10 These range from those focused on the near future of around 2030 to longer range futurist projections covering the next 10,000 years (contrast Tegmark, Citation2017; Harari, Citation2017).
11 Numerous attempts to state the scope of effects in terms of categories have been formulated. See, for example, WEF https://www.weforum.org/agenda/archive/fourth-industrial-revolution/ and also WEF Digital Transformation Initiative (DTI), initiated 2015: http://reports.weforum.org/digital-transformation/ Note there is also great scope for effects on agribusiness.
12 This is just one of the areas where various other technologies may come together: smartphones, blockchain, cryptocurrencies and so forth.
14 For example, Ford (Citation2015) argues that this time is different because machines have ceased to be tools and are now workers and the long rise or virtuous feedback loop between productivity, employment and wages has been broken, something partly illustrated by the increased inequality in the United States in particular and other countries in general in recent years. Disruptive technology is a system wide problem requiring careful restructuring in order to enable prosperity to continue for the many.
15 Amongst other things this creates additional context for recent debate over ‘bullshit jobs’. That is, capitalism’s capacity to create meaningless jobs that seem to serve no obvious purpose and that the worker knows need not exist (the organization would seemingly continue without it). These create new scope for alienation and are an odd mirror of the over-employment problem that existed in command economies such as China (e.g., the person whose job it was to sit by the door and guard the key). David Graeber (Citation2018) initiated the current discourse in an article in Strike in 2013; a UK YouGov poll in 2015 found that 35 per cent of employees think their job is meaningless and 33 per cent experience no personal satisfaction in doing it. In any case, there is a normative issue of social value related to CEOs, investment bankers, etc. despite the fact that they may be at low risk of displacement.
16 Note, though the movement is still generically referred to as accelerationism, Srnicek and Williams have dropped the term as potentially misleading, since the aim is to transform and in some ways slow human existence against the trends inherent in modernity.
17 The issue of ownership, in turn, raises longstanding issues regarding the nature of property and how these might also develop in a post-capitalist context (see Ireland & Meng, Citation2017).
18 Land’s early work is expressed in the postmodern idiom that had captured much of radical thought at that time. It is characterized by impenetrable self-referential verbiage, where no idea is completed or justified according to prevailing standards of reason and evidence (these are enemies), and whilst assertions abound, nothing is clearly explained in a way a reader might interpret as an attempt to be understood, nothing is definitively stated because nothing can be defined; one merely allows a stream of neologisms to accumulate and calls this an intervention. For example, ‘Since the history of thermodynamics is the history of technicizing commerce – of modernizing machines – any account that autonomizes science inevitably moralizes social change (into political theatre)’ (Land, Citation1995, p. 133). Land’s writing works as a kind of provocation, sentence by sentence and in a literary sense. He has an excellent turn of phrase, but from the point of view of everything his style rejects, it is pretentious bombast.
19 Later we argue that the fourth industrial revolution material has pre-empted debate by its capacity to quantify the future. This bleeds into the material published by and limited arguments of organizations like the TUC since they tend to rely on the data produced by the various sources associated with the fourth industrial revolution. They use these to position the significance of the issue, despite the fact that, as we shall argue, the numbers are dubious.
20 More prosaically, universal basic income (UBI) has its critics (see Fleetwood, Citation2014). UBI raises issues regarding the role of the state as a provider of services in relation to how basic income is provided, what it is spent on, and what it is intended to replace (there is a neoliberal variant of basic income where the state is hollowed out in favour of further privatisation).
21 For a specific argument at McKinsey regarding the scope for prediction offered in the context of AI see: https://www.mckinsey.com/business-functions/mckinsey-analytics/our-insights/the-economics-of-artificial-intelligence
22 Note, one must be careful not to misrepresent WEF and other contributors. The issue is emphasis and framing rather than lack of acknowledgement that many issues and considerations apply. For example, discussion of ethics and AI can be found at WEF’s site: https://www.weforum.org/agenda/2016/10/top-10-ethical-issues-in-artificial-intelligence/
23 For example, in December 2015 Professor Frey was interviewed by Raconteur for the headline piece in their Future of work supplement provided with The Times. For an archive of Professor Frey’s media activity see: https://www.oxfordmartin.ox.ac.uk/people/453
24 Autor (Citation2015), for example, is the main contrasting paper and that had 1,026 citations, whilst the OECD paper Arntz et al. (Citation2016) had 702 (see later).
25 In any case, this would likely replace one problem of atomism, regularity and closure with another (see Fleetwood, Citation2017). Autor, for example, favours production functions and more standard econometric analysis.
26 The use of the term ‘risk’ for the probability is perhaps intended to account for the issue of the ‘given’, but presupposes the efficacy of deriving a numeric probability for a conditional possibility (an issue ultimately of whether any kind of relevant distribution can reasonably be assumed to exist and can in some way be estimated or derived). Philosophy of mathematics and of analytical statistics has produced a great deal of critique of the problem.
27 The Hawksworth et al. (Citation2018), for example, repackage the transformative effects of a fourth industrial revolution into three successive (but partially overlapping) waves: 1. Algorithmic (affecting data driven-processing employment sectors in the early 2020s); 2. Augmentation (affecting robotics, warehousing and also more complex decision making tasks for data by the late 2020s); 3. Autonomous (extending to transport and construction, but building on 1 and 2 to affect all sectors to some degree by mid-2030s). They apply this to the same OECD dataset as other research and forecast gendered and education based displacement possibilities for current employment ranging from around 5 per cent initially to around 45 per cent in the mid-2030s for those with low education (affecting men more than women).
28 Workers of course are socialized by this. For example, the PEW Research Center’s The state of American jobs survey based on a sample of 5,006 people in 2016 found that 54 per cent of participants felt that retraining was essential to maintaining their employability and 72 per cent responded that it was the individual’s responsibility to seek out training.
Additional information
Notes on contributors
Jamie Morgan
Jamie Morgan is Professor of Economic Sociology at Leeds Beckett University. He co-edits the Real-World Economics Review with Edward Fullbrook. He has published widely in the fields of economics, political economy, philosophy, sociology and international politics. His recent books include Realist responses to post-human society: Ex machina (Ed. with I. Al-Amoudi, Routledge, 2018); Brexit and the political economy of fragmentation: Things fall apart (Ed. with H. Patomäki, Routledge, 2018); Trumponomics: Causes and consequences (Ed. with E. Fullbrook, College Publications, 2017); What is neoclassical economics? (Ed, Routledge, 2015); and Piketty’s capital in the twenty-first century (Ed. with E. Fullbrook, College Publications, 2014).