More VERY Dark Thoughts on Automation

More VERY Dark Thoughts on Automation

In March of this year, I made this post, http://writingsoftclaitken.blogspot.com.au/2016/03/an-extremely-pessimistic-prediction.html, inspired (or rather, intrigued and terrified) by three articles in the MIT Technology Review by a man named David Rotman about the dangers of the immense job loss from increased automation over the next few decades and the highly probable concomitant acceleration of the inequality that has been growing in the West since 1980, as well as a very polished Youtube video about cutting-edge robotic technologies being developed around the world and the millions of job losses these will inevitably bring about very soon in a whole range of professions – with the transport and retail sectors most in danger right now. After a brief meditation on the facts provided by these materials, I had come to an extremely dramatic conclusion: I had decided that there was genuine reason to fear the potential eventual development (in as soon as twenty or thirty years) of an entirely dystopian world: a world divided between an oligarchy of technological overlords, mostly composed of Silicon Valley geeks and their family and connections (perhaps along with other financial and political elites from before the transformed world), and an immense underclass of unemployed exclus and serfs living in poverty and disease in sprawling shanty towns and slums who spend their lives gazing up in primitive, superstitious awe at the magnificent, palatial cliffs on which sits the sparkling, porcelain-white technological utopia of the new oligarchs, who sit in porcelain-white rooms wearing Oculus Rifts, perpetually fed through tubes into their half-human, immortal cyborg bodies as they have fully realistic experiences of virtual sex with beautiful women, winning the Olympics to the cheers and adulation of millions of adoring fans, and living as heroic, manly Roman emperors and majestic warriors, venerated as Gods everywhere they go – this whole oneiric, atemporal, Huxleyesque fantasy sustained by the endless toiling of thousands of self-sufficient, self-repairing robot slaves, ceaselessly working to enhance the virtual reality technology for their human masters as the human ‘masters’ themselves sink further and further into the neverending ecstasy of their Experience Machines, wandering endlessly through the narcotic hinterland of the  H Y P E R R E A L like a delirious backpacker stumbling about in the blazing desert sun (strings of saliva hanging from his cracked lips and straggly beard), hallucinating oceans as he slowly dies of dehydration.

Immediately after I wrote this brief post, however, I suddenly realised (or so I thought) that things couldn’t possibly quite become so dystopian, because, at the point that millions of people start to get laid off as a result of the corporate automation revolution, spending power in the economy will dramatically decline, and this will massively reduce corporate profits by reducing consumption, and thereby give all the major business elites huge incentive to assist the state in implementing a Universal Basic Income in order to maintain consumer capitalism by giving the millions of newly unemployed people the means to continue buying the products that the corporations want people to buy (with unprecedentedly massive advertising campaigns (exploiting new technologies) likely implemented to try to get these unemployed people to spend as much as their Basic Income as possible on luxury goods). Even though this would still be a dystopian world, I figured that it would be a dystopian world that would allow people to live materially comfortable lives, and hopefully pursue worthy goals. It was even possible, I thought, that it would turn out really well in some ways: without any vocational burdens to worry about, perhaps we might even get thousands more people practising science and philosophy! Of course, the more likely possibility is that the vast majority of the unemployed people would spend their lives playing increasingly advanced, increasingly immerse, increasingly real video games – gradually adopting Functionalism about space and time, and discarding the idea of one true reality (which is a very disturbing idea). And if pretty much everyone did start spending all their time entertaining themselves with virtual-reality video-game technology, it may well be that the entire Western intellectual and literary canon disappears, and Western science is left to either machine intelligence or a select few elites. But while this is a very disturbing thought, at least such a world would be a world without poverty, since the UBI would prevent it.

It might have been after I had this series of thoughts (but I’m not sure) that I learned that Universal Basic Income is currently actually a popular idea among Silicon Valley Libertarians, and for the reason I came to on my own (to maintain consumer capitalism in the face of colossal unemployment). The Silicon Valley supporters of UBI do indeed believe that unemployment will dramatically rise over the next few decades, and getting the state to use its corporate tax revenues to give all these unemployed people a stipend for consumption will be a good way of keeping their profits high as earned incomes slide.

Anyhow, once I had this epiphany, my worries about our automated future decreased somewhat. Of course, on top of worries about mass descent into H Y P E R R E A L I T Y, I did still have reason to fear the continuing rise in inequality, and the growing power of the major tech companies today, which I realised could have the potential, once the mass robot manufacture begins, to become nearly as powerful as states, or even supplant nation states entirely (which I knew is another thing that many Silicon Valley Libertarians dream about).

The reason I am updating my old post is that I had some new thoughts on these issues a couple of days ago, and I have decided that there are big reasons to worry about this last problem (as well as a few very good reasons to think that the Universal Basic Income, if really given to a large chunk of the population who were fully idle (not generating new economic activity by participating in production and not acting as entrepreneurs, but simply consuming), could cause dangerous inflation).

Here are my new thoughts:

Contra Marx (with his false "labour theory of value"), it is extremely judicious, from a profit-making perspective, for a company or corporation to introduce efficient automation to replace the labour of wage-draining, inefficient, lazy humans. The reasons are fairly straightforward: machines, computers and robots only have to be purchased once (they don’t have to to be paid wages), they can most likely work all day and all night 7 days a week 52 weeks a year (no weekend breaks, no holiday leave and no parental leave), and, whilst they may need maintenance or occasional repair, they don’t get sick and they don’t suffer variable performance because of emotions. The initial investment in such technologies will almost certainly cost a lot, and your electricity bill may be higher (certainly with current automation), but these costs pale in insignificance to the continued costs of paying wages for humans. It is thus no surprise, with recent technological developments, that factories are no longer the only type of productive enterprise replacing workers with robots (a process that has been going on since at least the 1980s); now the retail and hospitality industries are being automated, and soon the transport industry will be also.

First, a discussion of automation in the retail and hospitality industries.

As anyone who has left their house in the past couple of years would know, supermarkets, airlines and now fast food chains have been, for the last couple of years, steadily replacing vast numbers of employees at checkouts and counters with shiny, touch-screen, easy-to-use machines that can process the same tasks instead. This development makes extremely good business sense because the massive reduction in wage payments and nontrivial increase in efficiency delivered to the companies by this change far outweighs the temporary cost of installing the technology, the cost of the more eerie and inhuman atmosphere of the outlets, and the cost of alienating older customers who struggle with the new technology. Because of the massive benefit of this transformation, it is essentially certain that this trend of retail and hospitality automation will continue to grow bigger and bigger over the next few years. Indeed, I think it extremely likely that – certainly within the next decade, if not within the next five years – we will see even relatively small companies deciding to invest in order-processing computers in order to reduce the number of staff they have to hire. There are already robots that could do the same job as a barista, and it’s not hard to imagine mobile robots soon being developed that can do menial work like sorting items at retail stores (putting clothes back on racks, stacking video games and books, and so on), or even waiting tables (though the majority of cafes, if they want to automate, will probably just install a few machines like the McDonalds ones, rather than getting robot waiters (which will probably by seen as an impractical gimmick until the robots become highly proficient at understanding natural language)). Obviously, most counter- and till-jobs could be replaced quite easily by payment machines.

Of course, many local, artisanal and indie businesses, restaurants and cafes will likely resist the movement in order to preserve a warm and human atmosphere which will distinguish them from the big franchises and chains. And naturally, as I suggested before, most people do prefer talking to a real human to touching a screen, which might tempt one to the conclusion that this automation of the retail and hospitality industries will be necessarily constrained – but I personally highly doubt that the retail industry’s automation will be significantly constrained. As I said, for a manager or business owner, as long as the number of customers coming in your door isn't too adversely affected by the cold, technologised atmosphere in your store, the reduced wage payments and improved efficiency enabled by the new technology are going to significantly increase your overall profit margin. Perhaps if we got to the point where shopping malls were full of touch-screen computers, and the average store had only one or two humans milling about (there would, of course, need to be a few security guards around to prevent shoplifting), the massively increased coldness of the overall atmosphere of such malls would help accelerate their decline – helping to complete the transition, begun more than a decade ago now, from real-world shopping to online shopping. In fact, it doesn't seem completely out of the question that shopping malls might completely die in the not too distant future, with probably a significant number of retail companies going under in the process (the ones slower to develop a major online presence). But, even if online shopping does just completely take over, the constant is that millions of retail jobs will be disappearing over the next few years.  

The transport industry is an industry in a similar position to the retail industry, and workers within it probably have even worse prospects. Indeed, it seems likely that the transport industry faces cataclysmic job losses within only the next three or four years. With highly reliable self-driving cars now already on the market, it will only be a matter of time before truck and freight companies start replacing unreliable, sleep-requiring, holiday-breaking, money-draining truckies with computers (as long as governments don’t impose laws to prevent this). Similarly, human taxi services will very likely become totally unviable once self-driving taxis become cheap. Indeed, just as pilots no longer fly planes (now they only control take-off and landing), it may be only a matter of time before only a tiny number of humans drive cars at all.

The print media industry is, of course, an example of an industry that has already been destroyed by technological change, and I believe it is also an example of an industry that will continue to be hurt by automation, now that computer programs are capable of writing articles of not awful quality (look this up). I think people doing journalism degrees at university in the current age are extremely, extremely imprudent, because there truly are almost no jobs available now, and the openings will be even slimmer in a few years’ time. Even the most esteemed publications in the world (like The New York Times) don’t have anywhere near the profit margin they had when they could sell millions of paper copies, because far fewer people bother to subscribe to the online edition than subscribed to the paper edition when there was no accessible content otherwise (also, in the case of The New York Times and The New Yorker, one can circumvent the paywall simply by going into incognito mode on Google Chrome). And while the popular media organisations that were created in this new era (Buzzfeed, The Huffington Post, etc) aren’t facing any new threats, it is very likely that these organisations will gradually shed more writing staff as automated-writing technology improves (especially Buzzfeed (one would think), given how little creativity and intelligence is required to create a listicle).

Furthermore, investigative journalism is gravely moribund – in fact, gasping for its last breaths on life support – and serious, non-investigative journalism has been largely democratised (people like me try to do it, without any expectation of remuneration).

There’s no doubt that old publications like The New Yorker, The New York Times, The Washington Post, The Wall St Journal, The Guardian, The Sun, The Daily Mirror and The Daily Telegraph will be around for several years, and possibly decades, to come. There are also more minor noteworthy publications managing to get by on donations (one very good one is the left-wing journal, The Baffler, and another is the science and philosophy mag, Aeon). But these publications are few and far between, and their subsistence is dependent on the sustenance of a literary culture that is, I believe, largely atrophying, largely due to technological changes (in particular, the creation of the smartphone, which chews up the majority of the leisure time of the vast majority of people under the age of 35, and the majority of the leisure time of the majority of people under 60, leaving, for most of those afflicted, almost no time for reading).

As for the other sections of the media industry (TV, film, music, video games, radio, advertising), they are in less immediate danger, and, except for the music industry, technological change has hardly put a dent in their current respective business models and profits. The reason the music industry is the exception is, of course, that the increasing dominance of streaming apps like Spotify at the expense of CD and record sales has created a situation where only the most successful musicians are able to even make a living without performing live, which has probably jeopardised the career prospects of literally tens of thousands of young musicians. But even though the music industry is somewhat of an exception, it is also true that far fewer people are watching live TV than only a few years ago, and far fewer are bothering to go to the cinema. Meanwhile, though radio seems like a true technological stalwart (as permanent as the wheel), it may be that it, too, suffers immensely over the next few decades (especially if people stop driving). And while one might be tempted to think that people in the business of creating these various forms of entertainment couldn’t have their jobs automated any time soon (not until artificial intelligence becomes genuinely creative), the date that this does begin to happen may not be as distant as we think, since there are even now programs that can write original music of passable quality and compose original poetry that is better than that of the average untutored human (look this up). Of course, it is very likely that consumers will still value ‘raw’ human creativity over machine-produced entertainment – at least for the time being, until the intellectual productions of machines begin to convince people that they are genuinely creative. Nevertheless, the threat remains.

Abstracting from any one particular industry, one sees an immense threat from automation to the several classes of jobs that involve essentially zero physical effort and don’t require any great cognitive skill, ingenuity or creativity, or those that require serious physical effort but don’t require any complicated physical performance. That is to say, all those jobs that might be fairly assigned the descriptors menial, administrative, secretarial or bureaucratic (involving simple procedures of searching, filing, tabling, sorting, organising, accounting and calculating), all reception- and call-centre- type jobs, and all manual-labour jobs that don’t involve any versatility of performance, are extremely vulnerable to automation within the next decade and a half. Truth be told, there are even many scientists vulnerable to automation within the next decade and a half – in particular, those chemists and bio-chemists who spend most of their time in laboratories.

Of course, if we manage to create machines or robots with true general intelligence within the next few decades (“AI-complete” machines, with capacities directly analogous to those of human thought, expression and ‘creativity’), these robots could, by definition, do anything a human can do, including putting forward revolutionary scientific theories about the nature of the cosmos, the nature of the mind, the nature of time and space, or whatever else, as well as philosophising about all kinds of deep things, and creating great works of art and literature (and doing historical analysis, and so on). They could thus replace humans in any job at all.

Now, I’m not saying that we will develop “AI-complete” robots within the next few decades – in fact, I think “AI-complete” robots (and I believe that the machines would have to be embodied, and to interact fully with the world, to be intelligent and conscious in the way we are) – are extremely far off indeed. Not just that: I think that the research and engineering that’s currently going on at Google is not even the right kind for general intelligence (Deep Learning is not a path to anything like human cognition or consciousness – though, to be fair, I don’t think anyone thinks that it is (in itself)).

But, as I’ve explained, this fact about the difficulty of reaching general intelligence will not in any way impede the massive juggernaut of job automation that will crush much of humanity over the next couple of decades.

Let us now focus on what will happen to companies like Apple and Google over the next few decades, and consider why there might be reasons to seriously worry about the possible trajectory of these companies, and the power they will wield over humanity.

In order to facilitate this focus, we will imagine a world where massive job losses have occurred in the retail and hospitality industries, in the transport industry, and, increasingly, in all those jobs that don’t require any great skill or ingenuity, as well as significant job losses in the media industry. Automation is ubiquitous in this world, and the vast majority of people still working work alongside machine intelligence (mostly non-embodied, some robotic). We are assuming that some kind of functioning representative democracy still persists (despite the fact that democracy has been increasingly corroded in the West since the beginning of the Neoliberal era, and is most decrepit in the US – where the work of the Princeton political scientist Martin Gilens seems to suggest that the bottom 70% of the population has essentially no say in government policy at all (due to the effects of money, donations, Super-Pacs, lobbying and the corporate media, all very skilfully synthesised by Thomas Ferguson in his Investment Theory of Party Competition) – and despite the fact that the decline of Western democracy seems set to continue). Since unemployment would have shot up over the years preceding the state of affairs we are currently in, with automation dominant, we can imagine that there would be significant popular pressure, as well as capitalist pressure (for the reasons outlined much earlier), to implement the UBI. So it is likely that a large number of people would be living on state welfare, with the majority of these people likely using this time to waste their lives on virtual reality gaming.

Now we face a problem: as I suggested earlier, giving a huge chunk of the population money to simply be indolent would probably create dangerous inflationary pressures, because, in order to avoid inflation, the government needs to spend money in a way that creates more productive activity in the economy, so that money doesn’t become disconnected from ‘value’. This is certainly not the case if everyone is lazing around (and spending their money on the same things – food, entertainment, accommodation and the like)). This massive risk of inflation is, in itself, obviously an extremely serious problem, and one that seems to me to admit of no obvious solution (it will just make things worse if you keep raising the amount doled out). And yet it is a problem that, I think, pales in insignificance to the set of problems I’m about to explicate, directly pertaining to the development of companies like Apple and Google.

What will happen when big multinational corporations begin massively automating their production? The main thing that will happen is that they will save a lot of money, and, while people are still consuming their products, they will make far bigger profits. Meanwhile, while they make bigger profits, the big Silicon Valley tech corporations will be working to develop machines and robots of ever greater capabilities – possibly even robots that can make other robots. This might seem fairly benign, but it is not, and I’m about to explain why.

If the UBI fails to create a satisfactory solution to the problem of unprecedented unemployment (and it almost certainly will fail because of the inflation problem), then inequality, already at levels without parallel in a hundred years, will be at levels never before seen in the history of humanity. The reason is that, at the same time as the big CEOs and Wall St banksters continue to increase their already obscene salaries and bonuses, large and increasing segments of the population will be sliding into abject poverty and squalor. This complete collapse of consumption and demand in the economy will inevitably bring down the entire political and economic system, and millions of people will be forced to resort to looting and violence just to survive, as cities descend into anarchy and infrastructure decays. The scenario is truly dystopian.

But the most scary fact of all is that corporations like Apple and Google won’t collapse along with the political and economic system, and for one simple reason: they will have massive infrastructure at their disposal, and they will have armies of slave workers, meaning that the collapse of the economic system will likely just spur them on to refashion themselves into the new states of this brave new world, with their own private infrastructure, healthcare, education and armies. Their robots will be able to repair and build other robots, and so their power will continue to increase, as technology continues to improve. The whole world could become polarised between an anarchist chaos of poverty, squalor and misery – a Hobbesian inferno too nightmarish to wish upon your worst enemy – and a few small Edenic oases of technological splendour and opulence. In this world, democracy would be entirely dead, and civilisation would be effectively dead, too. All would be in ruins, even as technology continues to improve within the magnificent, fortified palaces of the tech geeks who won out, and now get to spend eternity in virtual reality, wilfully oblivious to the hellscape they created. So the ultimate result of the automation, even with the UBI, would be exactly the one that I imagined originally, without the UBI.

So I think we should be worried.