‘Robots’ (or automated systems for manufacturing and distribution) seem to be on-trend amongst social scientists[1] (following on the heels of economists [especially those trying to peddle reports predicting futures] and tech evangelists), so here’s a deceptively simple question:
Do ‘robots’ (i.e. automation) replace/destroy jobs?
Lots of coverage across various media will give you an answer that is pretty much a categorical ‘yes’. In specific industries this is true, up to a point. It is true that jobs in specific manufacturing firms that bring in ‘robots’ do become redundant.
As this AP story on Yahoo (based on bits from Rand and some academic economists) suggests: “General Motors, for instance, now employs barely a third of the 600,000 workers it had in the 1970s. Yet it churns out more cars and trucks than ever.” However, as another bit of economic research from Deloitte attests that between 1992 & 2014 total employment rose by 23%. Indeed, academic economists Georg Graetz and Guy Michaels argue in a paper that “robots had no significant effect on total hours worked, [however] there is some evidence that they reduced the hours of both low-skilled and middle-skilled workers”.
So, what of this apparent paradox? Well, apparently, the jobs being replaced by automated processes are in specific sectors and at the same time those processes have removed jobs we have had a larger growth in other sectors, such as care (at least according to Deloitte). Bodily energy, in the wielding of a hammer etc. has been replaced by an automated machine process but various ‘caring’, ‘cognitive’ and ‘creative’ forms of work have massively grown in number [2].
I’m not an economist and neither can I necessarily verify the numbers and so on, but I am interested in the ways we use this information and these ideas of creation and destruction of jobs/work to tell stories about our social/economic future [3]. We are asked to buy into various forms of technological determinism. A particularly pungent example of this is the ways in which ‘algorithms‘, and the kinds of agency of computer system that word connotes, are said to have particular (mostly problematic or sinister) effects. Following Sally Wyatt’s excellent work on technology development we can point out some of the ways this determinism is functioning:
- We can be selectively descriptive in our technological determinism – explain and define specific processes as having an origin and impulse from particular technologies. Here, the story about GM above, is a kind of what Wyatt terms descriptive technological determinism.
- We can simply assume the technology leads in a particular direction as a kind of common sense. Many of us cannot imagine life without some of our technological supports: e.g. electric power, artificial light and so on. We already have things like this, so they will surely lead to other (faster, ‘smarter’, more sophisticated etc etc) things. Some robots have existed in manufacturing for some time, so that surely means we will have more, and perhaps in other parts of our life (automated home here we come!). This is a form of what Wyatt terms normative technological determinism.
- The most readily understood version of determinism in relation to technology is when we say: “x” will happen because of technology developments (in particular ways). This is rather common in relation to ‘robots’ – for instance Deloitte in another report on “the state of the State” suggest around 865k jobs will be lost to automation by 2030. This is what Wyatt calls justifactory technological determinism.
- We have methods for making claims about the world and these make particular kinds of normative and epistemological assumptions about technology development, innovation and use. In this way there are forms of what Wyatt calls methodological determinism (and maybe that’s part of what I’m doing here too).
Of course, this isn’t only about the rationale of future orientation – it is also about the kinds of imagination that rationale both draws upon and produces. If we tell stories about the apparent destruction of working time then we can also tell stories about how that may lead to an increase in leisure time – as JM Keynes famously argued in his Economic Possibilities for Our Grandchildren.
One of the more interesting arguments in the imagining of an automated future is what a given society should do with the apparent wealth of time freed up by the robots. The ‘universal basic income‘ is one such story – that everyone should be granted a share of the wealth of the productivity gains through the granting of a universal stipend that covers the basic cost of living. Here we enter the realm of (depending on your standpoint) attempting to think beyond capitalism – in the vein of ‘accelerationism‘ and so on (and, of course, there are critiques of this).
Like many forms of discussion about the development of technology, the gap in thinking about automation, I think, is the easy slip into affirming the stories we are told about technological development without questioning the assumptions and rationale of those stories and without attempting to tell our own. For example, while I am a great admirer of Bernard Stiegler’s work I think he falls into this trap when building parts of his narrative about an economy of contribution. It seems to me that there is perhaps some thinking to be done on what I’ve begun to call (for my own thinking purposes) the ‘automative imaginary’, the work it does and the ways it might be put into question.
We, geographers and social scientists, can and should collaborate in the study and development of automation, algorithms, robots, and so forth. However, we should retain our critical stance on all generalised attempts to declare that these technologies and systems have done this or that to young people, leisure time, jobs, and so on. This doesn’t preclude seeing how automating systems can be used to increase what gets measured as productivity, for example as the GM case discussed above does, or to determine behavior, for example as Natasha Schüll does in reference to Las Vegas slot machines. However, such instances usually depend upon a very specific context that we need to unpick. We should speak up for a critical and playful perspective to counter the rise of the apparent certainty that a particular version of modeling will ‘predict’ technological futures. We can and should be provocative [4].
Notes
- See for example the articles and upcoming conference sessions on robots in human geography: Social Geography II: Robots, robotic futures, digital\\human\\labour.
- It is interesting to note that for a while now, there has been a compelling counter narrative to the ‘robots are taking/will take our jobs’ story, which is a rather more familiar one: globalisation and post-Fordism. This story argue that rather than ‘re-shoring’ manufacturing to automated hi-tech plants, manufacturers are still paying for cheap labour ‘off shore’ and instituting ever leaner practices through ‘supplemental’ (to traditional human labour) robots. Likewise, Amazon, the apparent champion of robotising their processes, have a growing workforce, and have had for several years.
- It is depressingly predictable that these claims seem to rely upon the systematic dismissal or forgetting of work more-often-than-not identified as in some way female (such as care work), which has actually significantly grown, when focussing upon traditionally normative male work, like manual labour. Likewise, there is no accounting for the quality of work being generated by automation, which may be badly paid, more precarious and less regular.
- For example: Counting Sheep, The Museum of Contemporary Commodities, Uninvited Guests
