This essay is a response to the framing proposition, which states that the advent of technologies can narrow the urban-rural productivity gap by enabling younger generations in productive, fulfilling work without migrating to big cities. This proposition has been put forth in the context of China’s “small town youth” who are at the frontline of the country’s digital boom. However, this proposition disregards how labour relations are skewed as a result — workers do not have autonomy to decide tasks, negotiate payments or form worker collectives and are under the control,though fragmented, of the platform. In this context, the question of productivity gains as output is narrow, and disregards questions of worker well-being and more fundamentally, labour rights.
This essay examines the prompt by looking at the question of productivity in work generated on digital platforms. It asserts that the new digital economy needs a new definition and way of measuring productivity, which takes into consideration questions of labour well-being. The essay buttresses its arguments with analysis from India, which like China, has become a point of focus and expansion of the sharing economy, with Indian, American and Chinese platforms in the mix.
Understanding productivity in the digital age
The productivity gap, which is the measured difference in output between workers, or the gross domestic product per worker, is a cause of some debate globally. It is widely acknowledged that there has been a slowdown in productivity. There are three main reasons given for this perceived slowdown in productivity.
First is the claim by Robert Gorden <1> that today’s innovations do not compare in scale or impact to breakthroughs of the past such as the telephone or the internet, which had immediate, economy-wide implications. This view has been challenged on the grounds that new digital technologies are in their nascent phases.
Second, the OECD <2> argues that productivity has not slowed down; rather, it has become concentrated in a few firms that are more innovative and therefore more profitable.
While technology has, through gamified micro-tasking and AI labelling, as well as the gig economy, expanded opportunities for work-seekers, it has not necessarily enhanced well-being.
Finally, some claim that traditional measures of productivity have not accounted for the impact of the internet and digital technologies. Machines are outperforming humans in many industries <3> and there may not be a good way to capture this seachange. In this context, this essay examines the prompt, moving away from the perspective of productivity gains from tech-enabled work toward one that examines how to think about broader notions of well-being. While technology has, through gamified micro-tasking and AI labelling, as well as the gig economy, expanded opportunities for work-seekers, it has not necessarily enhanced well-being. These kinds of work are done often by those who are outside Indian megapolises, and those who lack avenues for employment elsewhere.
To unpack the complex intersection of productivity gains and technology enabled work, we examine two areas, mediated by digital technology, which have opened up for young people outside of Indian megapolises participation in the platform economy, and low-skilled digital micro-tasks (often gamified) such as object-labelling. This work, because it is mediated by technology and renders geographic location of the worker inconsequential, seems to provide opportunities for a large number of people and may appear to enhance productivity and opportunity. However, the nature of this work, and the protections and meaningfulness it offers to workers is under question.
There is a growing case, as Sen and Stiglitz <4> have suggested, for measuring well-being. If people’s lives will no longer be defined by the work they do, then is it important to measure contributions to GDP or should countries and firms be measuring labour health, education, political participation, charity contributions, and subjective wellbeing?
Labelling and micro-tasking: Little meaning beyond productivity
India is increasingly becoming a hot-bed for microtasks for the artificial intelligence (AI) industry. With the growth in the use of AI across sectors, data-labelling has become a key requirement. Simultaneously, the growth of the e-commerce sector has spawned demand for several labelling micro-tasks. Thousands of people scattered across tier II and tier III cities in India are engaged in labelling and annotating data.
The task of data-labelling requires relatively low levels of skill. Workers view data from cameras, sensors, emails and social media to highlight differences and similarities, or identify objects. Or they label objects on e-commerce websites (such as apparel) into proper categories. When this labelled data is fed into the algorithm, it can rightly infer the data, find patterns and learn from it over time. Data is used, for instance, for the self-driving car industry, to label images of road signs, traffic lights, and pedestrians. This data will be used to train the autonomous vehicles to navigate real life situations. Data labellers are the “construction workers” of the digital economy, putting together the pieces that hold together technological companies. <5>
The task of data-labelling requires relatively low levels of skill. Workers view data from cameras, sensors, emails and social media to highlight differences and similarities, or identify objects.
This industry, like the previous influx of business process outsourcing (BPO) work, can use India’s large, untrained workforce. Amazon’s MTurk was a popular way of crowd-sourcing micro-tasks in India, but restrictions were imposed for non-US based workers due to security concerns. As a response, Indian companies such as Playment, <6> iMerit and Infolks have been set up to cater to global clients and evolve into a hub for labelling and annotation work. Other companies like Squadrun provide labelling services by gamifying the labelling tasks to further engage those doing the work.
Data labelling companies, as mentioned above, tend to work out of smaller towns and cities, which are cheaper locations to set up these large centres as cost arbitrage is a significant element of the business model. The workforce is sourced locally, and often comes from families that make less than USD 100 per month. <7> Companies claim to pay labellers anywhere between USD 300 to USD 400 per month, adding significantly to household incomes.
There is little information about work conditions of data labellers in India. Even if one regards this work force as formalised, the quality of work, the nature of the work conditions, and the meaning workers derive from it raises concerns. The primary cause for worry is that these low-skilled jobs appear transient, which is that they only exist until machines themselves can be trained to label and annotate objects. Once that shift occurs, these hundreds and thousands of jobs could vanish and the people doing them would have to acquire new skills to stay relevant.
The primary cause for worry is that these low-skilled jobs appear transient, which is that they only exist until machines themselves can be trained to label and annotate objects.
This work has been defined as “ghost work” — invisible labour that powers technology platforms. <8> The labour force working to label data, is invisibilised, as technology companies tend to treat them as “code” and not real workers. This invisibilisation also keeps up the idea of technological magic, <9> a much-marketed myth which positions automation as a way of freeing up human time, but in reality, is fuelled by humans themselves. Perhaps, it is also why these jobs are kept out of big cities and are performed in far-off parts of the world such as rural India and China.
That such work able to provide opportunities for India’s youth, particularly in the rural and peri-urban areas is clear. However, how satisfying and sustainable these opportunities are, what opportunities they provide beyond transient gains in productivity and how well-being needs to be understood in context must be explored further.
Platform/gig economy work industry (ride sharing): Precarity and well-being
Ola and Uber both have recently launched “lite” <10> versions of their apps to reach out to the vast number of users and drivers in tier II and tier III cities in India. The apps are built to load on lower end phones, and use less than 1MB of data. In addition, both platforms are aggressively expanding their two-wheeler outreach in smaller cities, with investments in business such as Ola Bike, Bounce, Rapido and Yulu. Ola has also expanded to over a hundred cities in India and is becoming ubiquitous in state capitals across the country. Related to this national expansion of several technology platforms, a total of 1.3 million Indians joined the platform economy between October 2018 and March 2019.
As these platforms expand into smaller cities, there is an assumption that workers will be sourced from neighbouring towns and villages, coming into cities as shortterm migrants. A recent study shows that 30% of drivers <11> on ride-sharing platforms in Bangalore were agriculturalists or garment workers from neighbouring areas, pushed to the city due to agricultural distress. As in the case of labelling, the skills required to participate in the ride-sharing economy are limited. Increasingly, commercial driving licenses, which have a stricter criterion for approval, are not required for driving taxis.
Aapti’s empirical research shows <12> that workers lack job security and do not have social security in the form of insurance or health benefits from the platforms through which they work. Most workers in the platform economy do not consider this work to be a long-term option. They also must navigate complex and opaque contracts, keep impossible work hours to fulfil targets to unlock incentives. Workers experience arbitrary deactivation, surveillance and do not appear have any avenues for grievance redressal.
As in the case of labelling, the skills required to participate in the ride-sharing economy are limited. Increasingly, commercial driving licenses, which have a stricter criterion for approval, are not required for driving taxis.
Platforms experience attrition rates which can range from 40%-300% for some companies. Interestingly, half the workforce that is below the age of 23 years leaves these jobs in the first three months, either because of new opportunities, or because living in cities in these informal jobs becomes unviable.
While the gig economy does indeed add to productivity and opportunities available for those outside of big megapolises, the precarious nature of the employment raises questions about the meanings of well-being.
Implications: The well-being argument
As illustrated by the examples above, digitisation offers several opportunities to people, especially the technologically savvy, smartphone-using youth, so that they do not have to move to live in metropolitan cities. They can also find work closer to their homes in smaller towns and villages. However, the kind of work being mediated by technology raises several questions from the paradigm of well-being for the workforce, and the broader implications of these jobs.
While jobs in AI labelling and related micro-tasking indeed provide economic opportunity and enhance productivity by utilising idle time, their very nature — limited prospects for betterment and growth — and uncertainty complicate how to infer well-being. Similarly, in the driving platform economy, questions around hours, surveillance and emotional labour and more go to the core of inferring well-being. While new legislative efforts, such as the Code of Wages Bill, 2019 do attempt to include the platform economy and gig workers within their scope, we need to evaluate the broader set of concerns that arise in the tussle between enhanced productivity and opportunity on the one hand, and well-being on the other. <13>
While jobs in AI labelling and related micro-tasking indeed provide economic opportunity and enhance productivity by utilising idle time, their very nature — limited prospects for betterment and growth — and uncertainty complicate how to infer well-being.
The thrust of some of this innovation has been an entrepreneurisation of labour. While this may offer some advantages in terms of flexibility and choice, especially to those with some privileges such as social and economic capital, it may push those currently in the margins further out. Those who cannot afford the car or a smart phone that is a crucial requirement for entry into this sector, may be left at the margins of society, being unable to access these forms of employment. Women, especially, are significantly disadvantaged here.
Conclusion
We go back to Sen, Stiglitz and Fitoussi’s argument <14> that well-being is a better measure of growth and productivity and a focus should be on jobs that improve the quality of life by providing better health, education, social security and safety etc. It is crucial to better understand the aspirations of people and find opportunities for work that create a space for individuals to find meaning as opposed to performing jobs that enact transient ideas of productivity.
This is possible to do through investments in small and medium enterprises outside Indian metropolitan cities, encouraging the adoption of digital tools such that they are no longer on the margins of productivity, whatever the definition. It would also come from a thoughtful evolution of a legal framework to offer protections for those in the platform economy, and avenues to seek redress.
It is also crucial to move forward from static meanings of both productivity (increased GDP) and well-being to understand the platform economy in context. While there is freedom, flexibility and empowerment that come from participation on the platform economy, a focus on productivity alone may limit opportunities for those on the geographical margins in the long-term.
Endnotes
<1> Gordon, Robert. “Is U.S. Economic Growth Over? Faltering Innovation Confronts the Six Headwinds”, 2012.
<2> Andrews, Dan, Chiara Criscuolo, and Peter N Gal. “Frontier Firms, Technology Diffusion and Public Policy”. OECD Productivity Working Papers, December 2015.
<3> Mohan, Deepanshu. “India Is Very Much Part of the Global Productivity Slowdown”. The Wire. Accessed September 18, 2019.
<4> Fox, J., 2014. The Economics of Well-Being. Harvard Business Review 8 October 2014.
<5> Yuan, L., 2018. How Cheap Labor Drives China’s A.I. Ambitions. The New York Times 25 November 2018.
<6> Playment, in fact, states that it engages over 300,000 workers, and recognises about 25,000 of them, as “highly skilled”.
<7> Murali, A., A. Sen & J. PK, 2019. How India’s data labellers are powering the global AI race. FactorDaily 20 March 2019 .
<8> Chen, A., 2019. How Silicon Valley’s successes are fueled by an underclass of “ghost workers”. The Verge 13 May 2019.
<9> BBC News, n.d. The ‘ghost work’ powering tech magic - BBC Worklife. BBC News.
<10> ETtech.com, 2019. How Uber Lite is trying to reach a new set of potential riders in emerging markets - ETtech. ETtech.com 10 January 2019.
<11> Surie, A., & Sharma, L. V. (2019). Climate change, Agrarian distress, and the role of digital labour markets: evidence from Bengaluru, Karnataka. Decision, 46(2), 127-138.
<12> Gupta, S., 2019. Future of Workers: Building safe workplaces of the future. Medium 18 September 2019.
<13> Jha, S. & N. Alawadhi, 2019. Gig workers set to come under labour laws, get social security benefits. Business Standard 18 September 2019.
<14> Fox, J., 2014. The Economics of Well-Being. Harvard Business Review 8 October 2014.
The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content — blogs, longforms and interviews.