1. Introduction
Digital labor platforms have grown exponentially over the past decade and become a part of everyday life. User data has increasingly become a valuable strategic economic resource on these platforms, massive amounts of it being utilized for different economic activities, and the development of artificial intelligence (AI) and machine learning algorithms. With greater technological innovations and the availability of cloud infrastructure and computing services, digital labor platforms have gradually penetrated several sectors of the economy. But the cloud infrastructure itself is largely provided by a few large multinational companies, such as Alibaba, Alphabet (the parent company of Google), Amazon, Apple, Meta (the parent company of Facebook), Microsoft, and Tencent, located in the United States and China. More recently, the Covid-19 pandemic and the resulting prevalence of remote-working arrangements have further strengthened the position of these companies. Together, these developments have enabled Amazon, Apple, Alphabet (includes Google), Meta (includes Facebook), and Microsoft to emerge as the six biggest companies in the world in terms of market capitalization, each valued at over USD 1 trillion by the end of 2021.
This article will analyze the challenges that these multinational companies pose for developing countries, (re)shaping their economies and labor markets. Among other things, these companies have triggered a rise in online microtask and freelance platforms which are the latest manifestations of outsourcing services enabling businesses to adjust their workforce. The article will explore the implications of such an outsourcing model on the working conditions and skill development of workers, and its contribution, if any, to the productive development of the economy and society in developing countries. It will also examine how major platforms in the taxi and delivery sectors are capturing markets in developing countries, and the consequent disruptions in the labor markets with respect to decent working conditions.
2. Challenges for Developing Countries
Most large multinational companies providing cloud and computing services are predominantly platform businesses that facilitate interactions between users, applications, and service providers, though some also manufacture products. Companies such as Alphabet and Meta have also been investing in internet cables, with 80% of investments in new cables over the past few years coming from them. Simultaneously, U.S.-based companies have been building and controlling digital ecosystems that encompass technical, economic, social, and legal elements such as devices, platforms, users, developers, and payment systems, as well as legal contracts, rights, claims, and standards, among others. Such ecosystems lock in users, developers, businesses, competitors, governments, etc., resulting in new forms of “digital rentiership”.
The immense wealth concentrated in the hands of a few large multinational companies allows them to steer and manage innovations and shape infrastructure development in the digital economy. It also permits them to define the boundaries governing the digital economy and determine who can participate in it and under what terms. Furthermore, companies ensure adherence to these terms by developing mechanisms such as licensing of intellectual property rights and the technical frameworks through which cloud services are provided to users and developers.
As platform companies shape and govern the digital economy, they raise serious challenges for developing countries already facing major gaps in access to and availability of digital infrastructure. Such digital divides can be seen, for instance, in the internet connectivity of rural and urban areas of developing countries, where the rural connectivity rate is 34%, compared with 72% in urban areas. This gap has almost disappeared in developed countries. Even developing countries well known for their IT-enabled and software services, such as India, lag behind in internet connectivity and speed. Due to gaps in access to reliable digital infrastructure, domestic platform companies in developing countries face challenges in setting up and growing their businesses, as well as in fostering innovations. In contrast, big multinational companies are able to enter these markets and rapidly exert their dominance as they invest in and shape infrastructure development in these economies. In many African countries, for example, Facebook (now Meta) provides its Free Basics internet service and serves as a de facto e-commerce platform as well as a social media platform, and in the process is able to extract data. The dominance of few such companies from the global North risks exacerbating digital inequalities between countries.
Globally, the availability of cloud computing and cloud infrastructure services has led to a five-fold increase in the number of digital labor platforms (including web- and location-based platforms in the taxi and delivery sectors) over the past decade, but over a quarter (29%) of these are concentrated in just one country, namely, the United States. Faced with accelerating unemployment, especially since the 2008 global financial crisis, governments in developing countries have more recently begun to embrace such platforms and encourage their growth, with the hope of creating gainful employment opportunities. This has spurred massive government investments in building digital infrastructure and developing digital skills through training programs.
In turn, the expansion of the digital economy has led to two important developments in labor markets in developing countries, which are quite worrying. The first is the rise of online microtask and freelance platforms which offer firms and individuals a new means of outsourcing work, and the second is the emergence of new types of business process outsourcing (BPO) companies which cater to the needs of the digital economy. Both have crucial implications for workers in these economies.
The expansion of the digital economy has led to two important, but worrying developments in labor markets in developing countries. The first is the rise of online microtask and freelance platforms which offer firms and individuals a new means of outsourcing work, and the second is the emergence of new types of business process outsourcing (BPO) companies which cater to the needs of the digital economy.
Online microtask and freelance platforms: Developing countries are among the largest suppliers of labor to online microtask and freelance platforms, according to the Online Labour Index (OLI). This index, which provides an online gig economy equivalent of conventional labor market statistics, puts Asian countries such as India, Pakistan, Bangladesh, and the Philippines among the top 10 in this list. Work on these platforms range from short-term low-skilled microtasks geared towards AI and machine learning (data labelling, image annotation, content moderation, etc.) and promotion of products and services (content access, market research and reviews, surveys and experiments, etc.), to high-skilled tasks such as computer programming, web development, product design, translation, and legal and accounting services.
This new outsourcing model enables firms to access workers with varied skills to complete high-skilled as well as low-skilled tasks at a faster pace and lower price than if the same tasks were performed within the firm or in a sub-contracted firm. The transaction data on 200,000 projects collected and tracked on one of the largest online freelance platforms in the world between January and December 2019,1Transaction data for 2019 was obtained from one of the largest online freelance platforms to map the countries where the online work is performed across a range of occupations. The anonymized transaction data was obtained for the period January to December 2019 using the application programming interface; this is a sample of all projects on the platform, with a total volume of USD 135 million. shows that a disproportionate number of clients who outsource tasks are based in the Global North, including Australia, Canada, Germany, New Zealand, the United Kingdom, and the United States, while the work is predominantly performed by workers residing in the Global South. Firms prefer this model as it improves efficiency and reduces costs by allowing them to hire workers from low-wage locations. The median hourly wages in the United States (USD 25-35), for instance, are double that of India (USD 10-15), which alone accounts for 20% of the labor supplied to this market (resulting in cumulative earnings worth USD 26 million).
The outsourcing of work through online labor platforms has resulted in the creation of an invisible workforce that cleans, processes, and structures vast amounts of data into digital intelligence. Besides allowing firms to reduce costs, these “invisible” workers also help build data archives which can be monetized and used for various economic transactions and training machine learning algorithms for future automation. Their own working conditions, however, are often precarious.
Workers on these platforms are typically classified as self-employed or independent contractors. As a result, they do not receive any labor and social protection benefits. The average hourly earnings of workers in developing countries on online microtask and freelance platforms is USD 2.8, though about half the workers earn less than USD 1.4. These hourly earnings are not commensurate with their education levels, as most workers on these platforms are well-educated with a graduate or a postgraduate degree. There are also huge disparities in earnings as workers residing in developing countries earn 60% less than those in developed countries, after controlling for all variable characteristics. The low earnings could be due to the nature of tasks performed, but they are also likely the result of under-bidding or performing tasks without charging the client, as workers compete with one another in a global marketplace. Especially for workers in developing countries, this has the potential to lead to a race to the bottom, with implications for their current wages and future earnings.
Based on data-driven business models, taxi and delivery platforms are agile, asset-light, and can operate from remote locations. They make minimal investments in traditional capital assets, such as cars or warehouses, depending instead on data, skills, ideas, and physical assets provided by users, both clients and workers. This reduces entry barriers and allows platforms to enter new markets and expand rapidly through network effects.
In addition to the quality of work, there are also concerns about the content of tasks performed by highly educated workers in developing countries. About 54% of workers from developing countries on microtask platforms (AMT, Clickworker, Microworkers, CrowdFlower, Prolific Academic) have specialized in science, technology, engineering, and medicine (STEM) education. However, more than 80% of them perform tasks unrelated to their educational qualifications or skills, such as training AI and machine learning processes, promoting products and services such as on YouTube channels or websites, and writing fake reviews for tourist destinations, restaurants, hotels, or products. Some of these tasks are ethically questionable and none of them require any specific skill sets or training.
A global online survey conducted by the ILO revealed varying degrees of skills mismatch on freelance platforms, with about 40% of workers reporting not finding tasks related to their skills.2The survey was conducted among workers on two major freelance platforms, Freelancer and Upwork, between October 2019 and January 2020. While the two platforms operate globally, this survey was limited to 306 workers from developing countries in Africa, the Arab states, Asia, Eastern Europe, and Latin America. About half of the surveyed workers who had received an education in engineering and information technology (IT) said they found tasks related to their specialization, while the remaining had to take on professional services or activities unrelated to their skills. Similarly, about half of the workers who had a medical degree were performing business services or sales and marketing tasks. To an extent, workers with a degree in economics, finance, and accounting found tasks matching their skills, as 65% of them provided business services.
Business process outsourcing companies: The second trend that we observe with the rise of the digital economy is the proliferation of new types of BPO companies in developing countries. As AI is not fully developed, several Big Tech companies, such as Facebook, Google, and Microsoft, outsource content review and moderation, data annotation, image tagging, object labeling, and other tasks to BPO companies located in countries such as Kenya, India, and the Philippines. Typically, these tasks are outsourced as part of corporate social responsibility programs, with the ostensible objective of providing employment opportunities and making a positive social impact in developing countries. Some BPO companies which specialize in content moderation services boast that it is not only a business opportunity but also allows them to “act as a firewall or gatekeeper or a watchdog for the internet”. The use of social media, such as Facebook and WhatsApp messages, web chats, or emails, etc., to meet customer demand and provide real-time feedback has further strengthened the scope of these companies.
As in the case of microtask and freelance platforms, workers in BPO companies also face a mismatch of skills. Interviews conducted with workers in a content moderation BPO company in Bengaluru, India, as part of this study, revealed that about 90% of them are graduates or postgraduates with engineering and computer science degrees. Their tasks involve monitoring and moderating content on the web to remove offensive, obscene, false, or illegal content; preventing fraudulent practices on e-commerce platforms; moderating product reviews and advertisements; detecting fraud; and safeguarding copyright material and ensuring that there are no violations. Most of these tasks are unrelated to their education, thus throwing into question the utility of such services for the local economy and society. While such tasks are often perceived to be done by AI, in practice, they require human value judgment which is provided by BPO workers based primarily in developing countries or “invisible” workers on online labor platforms. This situation, in many ways, is similar to the outsourcing of waste (that can sometimes be toxic or hazardous) from developed to developing countries under the guise of recycling, which damages the local environment and the health of workers. In the internet economy, the toxic or hazardous content (such as disturbing images or videos) that is often generated by users in developed countries are sorted by workers in developing countries before it is posted on the internet, which has serious mental health implications for workers.
There are also concerns about workers’ career advancement in these jobs. Workers often mention that they pursue these jobs for lack of alternative employment opportunities. This, despite the fact that the Indian IT sector faces a shortage of qualified workers. It is interesting to observe that several IT-enabled service companies, such as Accenture, Genpact, and Cognizant, have also diversified and entered into the content moderation business, hiring university graduates to perform these tasks. The ongoing trend of tasks performed in BPO companies demonstrates the need to address mismatch of skills and ensure that the expertise of IT professionals is used productively for bringing about economic development.
3. A Few Digital Labor Platforms Shaping the Gig Economy
Taxi and delivery platforms comprise some of the largest and most well-funded labor platform companies globally, and mediate work for a large number of workers. Based on data-driven business models, these platforms are agile, asset-light, and can operate from remote locations. They make minimal investments in traditional capital assets, such as cars or warehouses, depending instead on data, skills, ideas, and physical assets provided by users, both clients and workers. This reduces entry barriers and allows platforms to enter new markets and expand rapidly through network effects. Some of these platforms have had far-reaching social and economic impacts in many countries, at times severely disrupting long-standing traditional sectors. A case in point is U.S.-based Uber which has expanded at an unprecedented pace across 71 countries while investing little in vehicles but benefiting enormously from network effects. As these companies diversify and offer a wide range of services, often through acquisitions or mergers, they begin to dominate multiple markets, resulting in a concentration of market power. For instance, Gojek and Grab in Southeast Asia have expanded and diversified into a wide range of services such as payment, taxi, delivery, e-commerce, etc., while Yandex, Delivery Hero, and Uber have undertaken 29, 28, and 13 acquisitions, respectively. Many of these acquisitions involved local and/or regional competitors.
Such unprecedented expansion has been made possible by venture capital (VC) funding. A high rate of return from network effects, which can result in a “winner-take-all” situation, bolsters the confidence of investors who bet on platforms dominating the market. While platforms with access to VC investments face very few barriers to entry in new markets, their presence in turn impedes the entry of other companies that lack VC funding. The first-mover advantage of highly funded platforms allows them to establish technological leadership, achieve rapid network effects, lock in users, and increase the costs they face in switching to other platforms. As the network of users of the dominant platform grows, they can attract more investments and further strengthen their market position. A similar “virtuous loop” is created with regards to data. With more user data, platforms can achieve higher efficiency, expand their services, and consolidate their position, which in turn allows them to gain more users and additional data. The ability of a few companies to achieve such network effects, reinforcing their capacity to lock in users and continue gathering even more data, enables them to transform into data monopolies (“data-opolies”).
A recent report by the International Labour Organization (ILO) shows that among the platforms analyzed, taxi companies received the largest share of VC financing. At USD 25.2 billion, Uber’s total funding was nine times higher than the USD 2.6 billion received by 142 online platforms combined. Uber’s revenue of USD 10.7 billion was the highest among all taxi platforms globally in 2019, and 36 times higher than that generated by the freelance platform Upwork. However, despite large revenues, many platform companies remain unprofitable. Uber, for example, has remained unprofitable since its founding, and in 2020 it recorded a net loss of USD 6.8 billion. But access to VC funding allows these platforms to operate at a loss for long periods of time, exacerbating their disruptive effects on traditional sectors. This situation has implications for start-ups and other digital labor platforms emerging at a national level, especially in developing countries, given their limited access to finance.
Among location-based platforms, for example, Uber has been able to rapidly enter markets by subsidizing consumers and drivers through discounts and incentives that allowed it to “exploit artificial market power to subvert normal market dynamics”. In India, competition between Uber and the domestic platform company Ola has resulted in a so-called “taxi war” to dominate the market, with VC-backed reduction in fares, provision of incentives for drivers, and advertising initiatives. This has not only led to the demise of some traditional taxi companies and loss of incomes for drivers, but also has implications for app-based drivers who increasingly face challenging working conditions, as well as consumers who are locked in with the dominant companies.
A comparative analysis of taxi drivers operating in the traditional model with those using platforms, conducted in nine developing countries, reveals that, at the outset, app-based drivers earn between 22% (Ukraine) and 86% (Ghana) more than their traditional counterparts. Taxi drivers on platforms can earn these higher incomes largely because of the disruptions wrought in the traditional sector and due to the bonuses and incentives provided by the platform companies through VC funding. However, over time, drivers on platforms face an increase in work intensity due to algorithmic management. On an average, they work about 65 hours per week, and as much as 82 hours a week in some countries such as India. App-based drivers also lose the freedom and flexibility they previously enjoyed as drivers in the traditional model. Algorithmic management also makes their work more precarious. The ratings drivers receive play a key role in their access to work — drivers are matched with customers based on these ratings — and their continuation on the platforms. As such, a decision to cancel a particular ride has repercussions on their work, leading to lower ratings, fewer rides, deactivation of accounts, etc. This was reported by about 40% of the taxi drivers who were part of the ILO survey. In addition, drivers on platforms pay 20 to 25% of their earnings as commission fees to the platforms and do not receive any work-related or social protection benefits. In these ways, platforms have plunged taxi drivers into precarity and disrupted the traditional sector, affecting their livelihoods.
Over time, drivers on platforms face an increase in work intensity due to algorithmic management. App-based drivers also lose the freedom and flexibility they previously enjoyed as drivers in the traditional model. Algorithmic management also makes their work more precarious. The ratings drivers receive play a key role in their access to work — drivers are matched with customers based on these ratings — and their continuation on the platforms. As such, a decision to cancel a particular ride has repercussions on their work, leading to lower ratings, fewer rides, deactivation of accounts, etc.
4. Reconsidering Digital Labor Platforms: The Key Takeaways
Platforms in the taxi and delivery sectors are shaping not only consumer needs and behavior but also the business environment for both traditional businesses and new local start-ups. Traditional taxi drivers and small businesses face major challenges in competing with platforms, unprofitable as they are, because of the disruptions they are able to create in the sector, backed by VC resources. These platforms also limit the growth of new businesses and employment opportunities for workers.
While online microtask and freelance platforms and BPO companies provide some income-generating opportunities, the relevance of such employment for local economies and the future career prospects of workers are, at best, questionable. They are the outcome of a lack of strategic thinking on the part of the government when it comes to creating employment opportunities, especially in developing countries where the educated workforce is growing.
In developing countries, higher education, and especially technical education, is associated with better skills. It is believed that such skills would facilitate the easy entry of workers into formal employment with stable contracts and regular incomes. However, there is growing evidence that the lack of suitable employment opportunities is forcing many highly educated and skilled workers to work on platforms where tasks are incommensurate with their qualifications, in order to earn an income or complement existing incomes. In some countries, the educated workforce is also resorting to new kinds of tasks on freelance platforms, such as academic writing outsourced by students in the Global North. This not only raises moral and ethical concerns but also has no relevance for local economic development.
The current development trajectory of the digital economy has the potential to push highly educated and skilled workers into informality, forcing them to work in precarious conditions without the benefits that workers who do similar jobs in formal firms would enjoy. It also has strong implications for exacerbating inequalities among workers as well as countries. The dominance of a few platforms from a few countries risks undermining the growth of local platforms and widening the existing digital divide between developed and developing countries.
Finally, the concentration of wealth among a handful of platform companies also poses a major challenge for regulating the digital economy and ensuring sustainable development in developing countries. The digital economy is currently being shaped by an emerging ecosystem of investments, mergers, and acquisitions concentrated among a few companies. Against this backdrop, developing countries remain reliant on a few companies for digital infrastructure and services, which in turn limits their capacity to undertake digital regulations and promote national digital industries, and hampers local innovation. Furthermore, with dominant platforms controlling a disproportionate amount of financial and data resources, the growth of alternative platform business models that cater to the development priorities of developing countries and promote decent working conditions is seriously undermined.
The digital economy is currently being shaped by an emerging ecosystem of investments, mergers, and acquisitions concentrated among a few companies. Against this backdrop, developing countries remain reliant on a few companies for digital infrastructure and services, which in turn limits their capacity to undertake digital regulations and promote national digital industries, and hampers local innovation.
It has thus become vital for developing countries to reclaim the digital economy and shape it in ways that reflect their own development priorities rather than those of a few platform companies. The situation also calls for a careful consideration of how emerging technologies can be utilized to bring about a productive transformation using a highly qualified and skilled workforce in ways that contribute to economic development and improved rankings on the human development index. In this context, two priorities that are critical for developing countries relate to data and AI.
The intensifying debate on free flow of data and data localization has major implications for developing countries. The United Nations Conference on Trade and Development (UNCTAD) points to three major approaches to data governance that are currently prevalent: the control of data by the private sector, as is the case in the United States; the control of data by the government, as is the case in China; and the control of data based on fundamental rights, as is the case in the European Union. These varying approaches create a risk of fragmentation of the digital space, with developing countries under pressure to choose between these dominant forms of data governance even as large platforms push to create their own data ecosystems. In addition, some free trade agreements (FTAs) also protect cross-border data flows of platform companies and prevent governments from localizing data. For developing countries, data governance has emerged as an important factor shaping development trajectories. It not only has implications for workers and local businesses, but also for innovation, taxation, and competition. However, regulatory discussions on cross-border data flows at an international level continue to be at an impasse. UNCTAD has hence called for a new global institutional framework for data governance that would enable data flows with development objectives, under the auspices of the United Nations, the most inclusive international forum in terms of representation of countries.
Despite the continued debate on cross-border data flows, there has been a proliferation of laws that focus on data privacy issues. As of January 2021, 145 jurisdictions had enacted data privacy laws, and 23 bills were passed on the subject, many of them in developing countries. In addition, countries such as Barbados, Brazil, and Panama have either adopted laws inspired by Europe’s General Data Protection Regulation (GDPR) or are in the process of doing so, while countries like India, Indonesia, and Pakistan have introduced bills to update existing laws. China’s Personal Information Protection Law (PIPL), which came into effect in 2021, also incorporates some of the elements present in the GDPR. Such laws can potentially have important implications for the platform economy given the centrality of data, as well as for workers’ rights as these laws may apply to workers as data subjects irrespective of employment status.
Another important area in the context of the platform economy is AI. Platform companies utilize algorithmic management practices to match workers with clients; rate, evaluate, and monitor workers; assign tasks; deactivate users; and dynamically determine pricing. These practices have also been associated with perpetuating biases and discriminations based on gender, ethnicity, and geographical locations of workers. For businesses, the use of algorithms by platform companies can result in competition issues, and competition authorities in different parts of the world have started to focus on their antitrust implications. These algorithms tend to be opaque, and many FTAs, such as the Comprehensive and Progressive Agreement for Trans-Pacific Partnership or the EU-Mexico, contain provisions that deny transfer of or access to the source code of algorithms.
At a national level, some countries have started developing regulatory frameworks regarding AI, but this has not quite taken off among developing countries. China has been moving rapidly with regard to regulating AI. In 2022, it introduced the Administrative Provisions on Algorithmic Recommendations for Internet Information Services which applies to recommendation algorithms used by platform companies. Its provisions cover aspects such as transparent disclosure to users, ethical use of recommendation algorithms, protection of the rights of platform workers, as well as oversight, such as through establishing an algorithm classification and classification of security management system. Such developments in China are being closely watched and may play an important role in shaping the global regulatory landscape on AI.
Ensuring transparency and accountability of algorithms will be an important challenge for developing countries to overcome as AI further penetrates different sectors and impacts governance, businesses, and workers. The ILO’s independent Global Commission on the Future of Work has called for “adopting a ‘human-in-command’ approach to AI that ensures that the final decisions affecting work are taken by human beings”. As developing countries move towards greater reliance on the platform economy for achieving their development objectives, public policies will need to ensure that algorithms do not function in a black box and humans remain in command, and that the undesirable impacts of AI do not exacerbate existing biases and inequalities in the digital economy.
