Techno-optimists believe that the future of work will be increasingly automated, and that algorithms will process large amounts of information at breakneck speeds, leading us to a new world of easy work. However, technological development has not ended jobs, but has created a lot of low-end jobs. Behind the glossy façade, millions of workers are manually processing data bit by bit.
Behind the rapid growth of the digital economy, from driverless cars to image search, is the little-known work of the dark box – processing large amounts of data is not AI (artificial intelligence), but gig workers who take orders remotely over the Internet. The gap between the vast poor and the AI-powered pioneers of wealth is widening.
On the one hand, the system no longer creates promising new jobs or drives productivity progress;On the other hand, cheap human labor becomes the end of the labor intelligence, engaged in the most boring and repetitive outsourced work, and is not guaranteed to be formally employed.
We see an embryonic future in which all formal jobs are dismantled and capital continues to own and control the means of production, but no longer employs humans. Workers derive all sorts of low-skilled labor from the remains of old jobs, providing data for machines and barely making ends meet. The tools that are supposed to illuminate the human world are throwing many people into the new obscurantism caused by technology. British columnist Phil Jones has clearly and meticulously exposed the brutal truth about the future of workplace automation in The Post-Work Era.
The Post-Work Era: Labor in the Age of Platform Capitalism, translated by Phil Jones and Chen Guangxing, Shanghai Translation Publishing House, August 2023.
Original author: Phil Jones.
The data morphs to.
All kinds of alienation machines.
We live in an era of technological wonders. Today, machines surpass humans in playing chess, writing pop songs, and self-driving cars. Automated stores allow customers to shop and check out on their own and then leave, without having to manually check out. Machines are clearly trying to decipher our minds through microchips implanted in the human brain. The ideal country of silicon wafers has made a promise to take our terminally ill earth to Mars to achieve immortality, free mankind from boring labor, and reach the realm of some kind of god. It's a world of affluence, where everything has been solved, and convenience and luxury are the norm.
But the foundations of this world are questionable, and the inevitable propulsion of scientific progress to move society forward is only the dream of a few tech tycoons. Negative dytopia, or utopian negative thinking, haunts this illusion of a harmonious society under computer cybernetics, and beneath its glossy surface, this harmony is based on increasing oppression, surveillance, and atomization. Every global historical event, whether it is a financial crisis or a pandemic, only seems to accelerate our progress towards its center: a "contactless future".
There, we are encouraged to avoid others and stay at home, where our home is no longer just a personal space, but also our office, shopping mall, gym, doctor and entertainment venue. The Internet of Things covers our sleep, meetings, and heart rate, turning every phenomenon into data that is then fed back into our lives as optimization services, all delivered by a platform.
Outside the home, "smart cities" have only provided tighter surveillance, with vulnerable populations struggling to survive under biometric and facial recognition surveillance, just like risk profiles. Algorithms weave all the bodies, spaces, and mechanisms into a vast machine-aware web so dense that computational intelligence becomes so commonplace that people even ignore it. Through this imperceptible matrix of sensors, trackers and cameras, capital is able to acquire new types of wealth that are known and recognized.
From meteorology to biostatistics, from the microcosm to the macrocosm, more and more life is enslaved by data exchange. Data morphs into all sorts of alienated machines: self-driving cars have replaced taxi and truck drivers, algorithms have replaced the authority of managers, and algorithms diagnose cancer more accurately than any doctor.
A still from the British documentary Horizon: The Age of Big Data (2013).
However, this automated dream world is more of a fantasy than a reality. Behind the search engines, apps, and smart devices are the vast numbers of workers who are banished to the periphery of a globalized system, devoid of other options, forced to clean up data and oversee algorithms, and paid only a few cents. Facebook and Twitter appear to be able to remove violent content automatically and precisely, but it's not algorithms that determine what constitutes pornography or hate speech.
Facial recognition cameras appear to be able to automatically recognize a face in a crowd, or a self-driving truck. In fact, the magic of machine learning lies in the heavy labor of annotating data. Behind the cargo worship rituals in Silicon Valley is the heavy lifting of sifting through hate speech, labeling images, and showing algorithms how to identify a cat.
The book argues that it's the low-paying, physically and mentally harmful work that makes our digital lives possible, not algorithms. Jeff Bezos announced to the world at the opening of Amazon's Mechanical Turk that "you can think of it as micro-work, so that for a penny, you can hire someone to tell you if there's someone in a table."
Amazon Mechanical Turk is the first and by far the most famous of its kind. On this type of **, the work of training AI by tagging the people on ** lasts only one minute. Even longer working hours tend to take no more than an hour. Microwork** allows contractors to break down larger projects into extremely ephemeral jobs.
Contractors post these "artificial intelligence jobs" (hits) on **, which appear on the screens of thousands of workers, those known as turkers (mechanical Turk platform workers), who will scramble to complete the jobs one by one. The platform receives a 20% cut from each transaction. This work is carried out remotely, and these workers never meet each other, except for the **forum.
Mechanical Turk** is the archetype of 21st-century work, empowering capital and dispowering workers. There are now many competitors who follow suit, providing contractors from research to capital, from research to capital, with exciting clean data and cheap labor. These are profit-making brokers who use their labor to find what Mike Davis calls the "surplus population" — the population that has been excluded from the global economic ontology — to meet the needs of Big Tech in a piecemeal way.
The length of time a worker has been hired is simply the length of a task, and the worker fluctuates between employment and unemployment status, most likely working for countless companies in a single day. These** package this uncertainty as flexibility, thereby claiming to be a beneficial, forward-looking guardian of a new type of labor contract that they claim is tailor-made for a generation of workers who want greater "independence" in terms of security and decent pay.
However, the only beneficiaries of this employment model are contractors – often big tech companies such as Twitter, Facebook and Google – who can escape the responsibilities that come with more conventional hiring. The people who work for these ** are no longer classified as "workers" but as "freelancers", "independent contractors", and – perhaps most shockingly – "players" who have given up rights, rules, and last shred of bargaining power.
The brutal logic of platform capital is transforming an already rather bleak global labor market into a gray area of temporary and transient employment. But when we read about microwork, we think that this kind of data work is a completely new phenomenon. The assertive assertions about "human cloud", "human as a service", and "instant work" suggest that the old-fashioned world characterized by the workday has leapt into a brave new world characterized by "human-machine hybridization". Jeff Bezos uses "artificial artificial intelligence" to show that in a "new economy" of ** growth, there is a high-tech agreement between workers and algorithms.
A still from the British documentary Horizon: The Age of Big Data (2013).
If the book has one goal, it is to convince readers that microwork is not a boon for the Global South, but a further distortion of the global work crisis. Microwork is the sum of intensifying trends such as slow growth, proletarianization, and declining demand for labor among marginalized groups in countries like India, Venezuela, and Kenya.
So it is perhaps not surprising that in the long period since the collapse of the 2008 crash, the number of people on these lists has skyrocketed. While there is no exact number of workers in microwork globally, it is currently estimated that the number is around 20 million, with a large proportion living in the Global South, namely South America, East Asia and the Indian subcontinent.
Many of these workers are better educated but out of the formal labour market. Among workers in the Global North who are over-educated and underemployed, the number of micro-workers is also rising. In the UK, surveys show that up to 5% of working-age people use these at least once a week**. For these workers, microwork is mostly part-time to increase working hours and subsidize stagnant wages. However, for many people around the world, microwork is a full-time job. A survey by the International Labour Organization found that 36% of workers work seven days a week.
Judging by the number of users published by each platform, the number of people working on these ** is likely to be much higher than current estimates. In the past 10 years, the number of users of ClickWorker alone has grown to more than 2 million, and the number of users of small users like Appen has exceeded 1 million. If the workers who use these platforms are classified as employees, then these contracted companies are among the largest employers in the world today, behind only a handful of countries** and Walmart.
For proponents of the Washington Consensus, the growing number of people who survive on menial data work are clear beneficiaries of AI, this argument is a light rebuttal in the face of a steady stream of reports that automation will create a large number of victims. And as it turns out, the line between victims and beneficiaries is not so clear here. Call center workers, replaced by chatbots, and checkout staff, replaced by automated checkout stores, are most likely to drift adrift in the capital storms of the 21st century, thus being forced into the dreaded sanctuary of the ** mission.
Automated. The last mile".
Proponents of microwork will insist that there is still work. But as the average wage for workers doing tasks on Amazon Mechanical Turk** is less than $2 an hour shows, even if automation hasn't wiped out workers entirely, it has now pushed them to the brink of survival.
This is related to the second theme of the book. The surplus population has not been considered normal human beings for a long time, they are victims of the cruel policies of the state. And now they are being treated inhumanly in the experiments of the Silicon Valley elite. As Bezos describes Mechanical Turk** as "artificial artificial intelligence," workers are not treated as humans, but as computer infrastructure. Programmers typically interact with computers using application editing interfaces, which connect employers and workers.
However, on microwork**, employers interact with people pretending to be computers. Workers disappear into the great shadow of the machine, so those employers, especially the larger platform customers, are able to stick to their marketing strategy without any interference. Following the strategy of Facebook, Google, Amazon, and countless start-ups looking to access venture capital, their business models are unusually streamlined, relying little on high-risk areas of labor and completely relying on complex algorithms. They promise to complete a process that Karl Marx did in the 19th century, that is, labor will be replaced by science and technology as the central element of capital productivity.
While platforms are accelerating this process, one only needs look at Foxconn, or the Ceroririco tin mine, known as the "man-eating mine," to see that these promises have not yet been fulfilled. Platforms outsource labor without crediting it to the ledger and hide the presence of labor from users, investors, and customers, making it appear more technologically advanced than it actually is, no different from the data work that powers AI.
While data is the lifeblood of a platform, data production is not what we usually think. We can see the hardware of the iPhone and see the labor required to make it from its materiality. But we can neither see nor touch the data transmitted through the software of the Apple phone. We will never be forced to face the fact that data must also be produced;Data, though ethereal and elusive, is the result of human labor, just like hardware. The work done by the human hands and brain seems to be just the output of intelligent machines, but this is a huge misunderstanding.
A still from the British documentary Horizon: The Age of Big Data (2013).
This obsession with data – seeing only automated drones, not the people who classify it;Workers who only see social** and can't see filtered information – obscuring the real hiding place of automation, which is the growing number of workers who are out of their regular employment and taking on the task of training machine learning intermittently.
Like Marx once entered the factories of the 19th century, to enter the hideout of automation, the book must also delve into the dark depths of platform capitalism, which has rapidly developed into the dominant economic model in the 21st century. By 2019, Amazon, Facebook, Microsoft, Alphabet (Google) and Apple were among the top five most valuable companies in the world. Crucial to the rise of these companies is the enormous computing power they already have at their disposal.
As a digital infrastructure that provides users with a space to meet, socialize, transact and consume, the platform is able to obtain a large amount of personal data from our browsing, global positioning location, our conversations on social and what we say in front of Siri, Apple's smart software. The more data these platforms accumulate, the more information they provide to AI, and the more automation they can get.
However, if the "last mile" of automation is really close at hand, it will be a very long mile. Even if Silicon Valley makes its dream come true with some impossible leap, automating the Congolese mines from which it gets its copper, the Foxconn factory that assembles parts into computers, and the Uber taxi that cars use to learn to drive themselves, almost all of these technologies rely on data processing — tagging, classifying, and categorizing — tasks that still lack technical solutions.
Not only do algorithms need clean data from the start, but they also rely on continuous monitoring and improvement when they start running. As Lily Irani puts it, "In order to adapt to changing environments, it takes manpower to configure, calibrate and adjust automation technology, whether those changes are different shapes of products or birds flying into the factory".
A still from the British documentary Horizon: The Age of Big Data (2013).
The Silicon Valley dream of fully automated luxury capitalism may indeed be a dream, but it remains a haunting dream for the 21st century, born of crises and prolonged recessions, built on crumbling democracies, and sporadic by climate catastrophe and economic austerity. At the moment, this imaginary utopia and the real-life dystopia are dancing towards disaster in a strange combination. While the chatbot is making significant progress, the California wildfires are burning. While computers beat humans on the chessboard, millions of people are suffering from the strange symptoms of zoonotic diseases.
Humanity is unable to grasp the momentum of historical development and create a better world, and humanity is about to face a future where smart taxis drive silently in the wind and rain that rages all night. As extreme weather events and pandemics turn more and more people into refugees, prisoners, and economic exiles, the burden of life – the inability to buy goods in economic agents – is presented to Silicon Valley in the form of software, which decides to use or refuse to use it as it wants.
But these perceived overpopulations are fighting back on multiple fronts, and their struggles may help create a better world. Microworkers, without the help of others, are already making earnest efforts to organize against capital. A growing number of events show that platform workers and other dispossessed people are joining forces. This is the little bit of hope that is cautiously held at the end of the book.
The truth about microwork.
It seems absurd to want to dream of finding enjoyable, meaningful work in micro-work. The monotony of tasks, the lack of routine, and the double instability of income and occupation all show that microwork is contrary to Morris's vision of Eden. However, these poor qualities are not inherent in microwork per se, but are the result of wage relations depriving the work process of satisfaction and meaning.
In fact, the promise of microwork to workers – independent, flexible, leisurely work lives – is strikingly similar to the world of work proposed by Morris. In a wage society, this commitment is undermined because these qualities often mean poverty and instability. However, in a wage-free society, microwork offers a surprisingly attractive vision of how to organize working life, as if a "concrete utopia" is hidden in the ruins of our day.
Of particular interest is the wide range of tasks that workers are able to accomplish in a single day. Thinkers such as Morris and Marx believed that this diversity was essential to work that met the needs of society and the needs of individuals. Microwork follows the principles of flexibility and independence, not because these qualities are beneficial to workers, but because they help capital evade its responsibilities to the workforce.
In a wage society, working for 20 task-posters in a single day, taking on as many roles as possible in the economic system, has been shown not to provide independence and motivation for workers, but only to a monotonous and protracted struggle for survival. But in a post-scarcity world, wages and their associated labor sub-industries have disappeared, and it is entirely possible to diversify the small amount of work needed for the benefit of workers.
Microwork** can be done with a wide variety of tasks in a day, revealing in a twisted way that labor can be very diverse in a world of common prosperity. As Marx and Engels said:
In a communist society, where no one has a particular sphere of activity, every man can develop in any sector, and society regulates the whole of production, thus making it possible for me to do this today and that tomorrow as I wish, hunting in the morning, fishing in the afternoon, animal husbandry in the evening, and criticism after dinner, but this does not make me a hunter, fisherman, shepherd, or critic.
Micro-tasks that undermine personal development, in a brutally exploitative way, provide a twisted version of this vision. But microtasks can still prompt us to question the fact that under our current capitalist system, those so-called jobs are merely pathetic simulacrums of occupations. In a digital age where work can be divided into tiny units, microtasks require us to rethink career coherence.
Of course, the task of harming the human psyche to varying degrees cannot replace what Marx called the exhausting journey of earthly pursuits. But if we focus on its organizational form, rather than its content, we have the possibility of finding the flexibility of the work envisioned by writers such as Marx and Morris. While many jobs, such as engineers, medical professionals, and teachers, are difficult to compartmentalize in the same way as data jobs, the time required to complete them can still be distributed in a fairer and more reasonable way.
In fact, a wage-free society still needs people trained in a specific occupation, but this profession is neither its sole role nor its main daily activity, but one of many. One can be a doctor for a few hours in the morning, then a farmer, and write a ** in the afternoon. Because there are enough people with medical training, no one has to sacrifice a rich and meaningful life to play this role. As Goetz writes:
Some tasks would be unbearable if you were doing them all day every day, such as sorting mail, collecting garbage, cleaning and repairing, but if you gave them to everyone, it would only take 15 minutes a day for each person to do them, and such work would be just a short interval between many tasks. If a job requires only a few days in a year or a lifetime, such a job may even become a popular pastime and pastime, such as a certain agricultural or forestry job that already has this characteristic by now.
Microwork** allows us to imagine a new world in which not only the time for each task is greatly reduced, but also the time for work as a whole. These** minimize labor remuneration and cause workers to spend more time looking for work than actually getting it done. The inherently irrational nature of this system, which makes everyone dependent on wages but unable to make them available to all, is not only an inefficient use of labor resources, but also harms the physical and mental health of workers.
In this case, fewer hours of work means less income and thus less chance of survival. But in a society where survival is no longer dependent on income, in a society where all people have been given the material basis for all-round development, this problem does not exist. With the help of technology, we are freed from unnecessary hard work, and social labor is only a small part of our lives. Just like on microwork**, machine learning algorithms can be used to calculate and distribute existing work, but always prioritize everyone's free time and autonomy.
In addition to reducing the total amount of work, machine learning can be used to improve the quality of the worker experience. The feedback infrastructure of big data can be used to assign work, but in a way that not only based on worker performance, but also takes into account worker preferences. But unlike the data factory mentioned in the previous chapter, where machine learning algorithms determine which workers complete which tasks based on their past performance, the algorithm prioritizes the happiness and interests of each worker when assigning work.
Gamification of work is possible.
Finally, gamification of work doesn't have to be an oppressive tactic either. If we set aside the economic purpose of motivating the managerial elite to adopt gamification strategies, we are likely to admit that even today's oppressive form of gamification of work embodies a desire to make work more interesting and enjoyable. In a world without wages, the joyful labor promised by companies such as Playment (data labeling platforms) will be realized, and labor will become fun and games rather than productive gains.
Work gamification is no longer used to regulate, rationalize, and monitor work processes, but to reduce monotonous drudgery. Machines will no longer be used to save labor costs, force workers to work harder, and make skilled workers do unskilled hard work, but will make work more interesting. To some extent, evaluation and scoring systems may replace wages as a means of incentivizing proper labor. Rather than deciding who will be paid and who will survive, the scoring system is a soft form of encouragement, a playful way of competition, and a game of oppressive hard work.
Edward Bellamy, in his 1888 Gilded Age's Looking Back, describes the socialist utopia of the year 2000, where wages disappeared along with money, inequality and war. Describes a system of moral and honor-motivated motivators to work, rather than brutal economic pressures to force people to work. There are some similar incentives in the public scoring system on microwork**. But while rewards such as Mechanical Turk's "master" status allow workers to secure better-paying work, in a post-scarcity economy, similar honors are instead used to confer outstanding status and motivate people to do monotonous or hard work.
The platform economy and its promised ideal state of silicon wafers are not only the source of the myth of the regeneration of capital, but also a social laboratory in which we can find amazing answers to the questions that have long plagued the socialist ideal. The question of when to use decision-making algorithms (and when not to) is unappealing to a platform economy that indulges in digital solutions.
Similarly, little attention has been paid to how job flexibility can really benefit workers, or how platforms can provide new impetus for jobs beyond wages. These are still the problems that any socialist conception must face.
As the previous chapter shows, no social ideal can succeed without the inclusion of the growing number of so-called surpluses. But to inspire a movement that can bring about a better world will require a vision of freedom and the necessities of life, one that is not only credible, but radically different from the current capitalist system, which can only bring more and more suffering and the collapse of the planet. For most people, the present is still tolerable, so much so that the future is still unimaginable.
Soon, an unimaginable future could turn into an unlivable one. In the face of climate catastrophe, capitalism has come up with only a techno-solutionist + accumulation + death-cult scheme. As if to demonstrate the destructive nature of capitalism, facial recognition cameras are made to arrest billions of people displaced by climate catastrophes, not to save them;A chatbot can only say its clichés meaninglessly while the earth burns.
In stark contrast to this failure of imagination is the imagination of Silicon Valley to profit from the victim population of the capitalist system, which has designed a form of work that is only slightly better than a life of total unemployment. Microwork points to a future where workers are primarily responsible for generating data and automating their work to eliminate it. But for the same reason, microwork can also point to a world without pay, where work is less important in our lives and we have more choices about when and what we work.
In the words of the historian Thompson, this ideal world does not rise on time like the sun, it must be made. The light of the world is becoming stronger and stronger in the midst of the struggles waged by a growing number of vulnerable groups around the world. At the darkest moment in history, the dawn of a new world began to flicker.
The content of this article is excerpted from Post-Work Era with permission from the publisher. Original author: Phil Jones. Excerpt: Lotus;Editor: Wang Han;Proofreading of the introductory part: Yang Xuli. Welcome to the circle of friends. At the end of the article, there is an advertisement for "The Scale of Time: 20 Years of the Beijing News's Good Book of the Year".