The urgent need for regulating global ghost work

People work on computers at the offices of online retailer Jumia in Nigeria's commercial capital, Lagos.

Technological innovation has long promised to lift the world’s poor toward prosperity, and the digital economy is no exception. The World Bank and the International Monetary Fund have adopted Silicon Valley-inspired narratives to argue that digital innovation holds immense promise for millions of workers worldwide to find work, but for laborers in the digital economy, these promises have been illusory.  

Working through freelancing platforms, digital laborers in Africa are filling vacancies for data-labelling and content moderation jobs posted by North American and European companies. While some local digital entrepreneurs have found fortune, African platform workers taking part in the digital gold rush face economic insecurity, unpredictable working conditions, and limited bargaining power.

Policymakers have failed to be proactively mitigate the entrenchment of global economic divides in the digital economy, and this oversight has left digital platform workers without basic labor protections, such as minimum wages, safety regulations, and clarity regarding taxation regimes. This unchecked labor market, in which workers perform tasks ranging from deskilled microwork to complex technical projects, has ramifications far beyond abuse of worker rights. The decisions made by data workers in Africa and elsewhere, who are responsible for data labeling and content moderation decisions on global platforms, feed back into and shape the algorithms internet users around the world interact with every day.  

The African century

Since the start of the 2010s, global leaders and local politicians presented Africa as the new frontier in the worldwide digital economy: a place of untapped wealth to be unleashed through increased digital connectivity. Though internet penetration in sub-Saharan Africa lags behind the rest of the world, there are at least 270 million people using mobile internet, and the cost of data is decreasing. According to a report by Google and the International Finance Corporation, Africa’s internet economy will be worth $180 billion by 2025 and $712 billion by 2050. Jack Ma, the co-founder of Chinese tech giant Alibaba, argues that “Africa’s entrepreneurs will drive the next digital revolution.”

In recent years, the continent has experienced a boom in digital entrepreneurship thanks to huge investments in entrepreneurship programs, co-working hubs, innovation prizes, and ICT infrastructure. Capital is flooding into the region, with some $55 billion pledged as part of the African Development Bank’s Connect Africa Initiative. France’s development agency has pledged $76 million to its African start-up fund. The World Bank has launched its own technology accelerator, XL Africa, that helps start-ups attract capital up to $1.5 million. The Konza City technology park outside Nairobi will cost the Kenyan government an estimated $14.5 billion as part of its Kenya Vision 2030 plan. The Nigerian billionaire Tony Elumelu has pledged $100 million as part of an effort to fund the next 10,000 African entrepreneurs seeking to jump-start their businesses. One estimate put total African start-up funding in 2019 at $1.3 billion. With this influx of cash, Africa’s rapidly growing cities are adopting the branding of Silicon Valley. Lagos has the “Silicon Swamp,” Nairobi the “Silicon Savannah,” and Cape Town the “Silicon Cape.”

Notable African success stories have fueled enthusiasm for this digital gold rush. Branded Africa’s first tech unicorn, Jumia, the “Amazon of Africa,” became a darling of the tech press and the first African startup to trade on the New York Stock Exchange and briefly reached a $3 billion valuation. Innovations in digital financial services is improving access to banking and other financial services. Kobo360 and Lori are two examples of African start-ups investing in mobile on-demand services that have led to new and better-functioning markets.

Such developments have encouraged the idea that in this new digital age, where technological infrastructures are made easily and cheaply available, the barriers to entry for digital entrepreneurship are relatively low. According to this myth of digital universalism, anyone with an internet connection can become a successful digital entrepreneur.

The growth of ghost work

According to the mantra of the World Bank and local governments, the adoption of digital technology makes work available to anyone from anywhere—not just to a privileged few but to all workers, including those with little or limited education. Encouraged by these stories, millions of people are now partaking in digitally mediated work to transcend their local labor market constraints and access new economic opportunities. The internet, it is promised, will allow them to strike gold. The reality, however, is very different.

Like Africa’s first tech unicorn, which experienced a specular fall on the NYSE a few weeks after its debut, much of the hype surrounding the growth of Africa’s digital economy is too good to be true. Even the region’s best economic performers—South Africa, Nigeria, and Kenya—have internet penetration rates of only about 50, 60 and 80 percent, respectively. ICTs are struggling to level the differences between economic cores and peripheries, and many of the scalable digital enterprises on the continent are being dominated by Silicon Valley firms, with whom local entrepreneurs struggle to compete.

The discourse surrounding the rise of Africa’s digital economy hides the nature of the bulk of digital labor being done by millions across the continent. Western enterprises are profiting from the lucrative and cheaply-sourced labor input of African digital workers, referred to as “ghost workers.” These “ghost workers” power mobile phone apps, websites, and artificial intelligence systems we interact with daily. They are the new extractive commodity.

Web-based digital labor platforms have ushered in a new form of offshoring and outsourcing for North American corporations. According to Oxford’s newly launched Online Labour Index (OLI), U.S. employers are the largest users of online labor, followed by the United Kingdom, India, and Australia. These digital labor platforms permit the real-time hiring of services from a global pool of low-cost talent, ranging from IT design, copywriting, and routine clerical tasks. These “ghost” human workers, available to work at all hours, are essential in maintaining systems and services that are assumed to be automatic or artificially intelligent.

The type of labor performed by these digital workers is often grouped into two categories: “microwork” and “macrowork.” The term macrowork typically refers to more specialized, knowledge-intensive, and long-term projects, such as software development and design and writing and translation. Macrowork contracts can be found on freelancing platforms like Upwork, Freelancer, Fivver and PeoplePerHour. Microwork, on the other hand, refers to simple repetitive tasks that are mostly clerical and require little training and coordination. Such tasks include product categorization, processing, verifying and validating data, and content moderation (e.g., removing pornographic or violent images before they are uploaded on social media platforms). Popular microwork platforms in Africa, Latin America, and Southeast Asia include Amazon’s Mechanical Turk (or MTurk), Samasource, CrowdFlower, and Microworkers.

This labor spans the digital economy and has important feedback effects for online tools used by billions. Digital workers, picking up contracts advertised on the MTurk marketplace and other platforms, can spend hours labelling pictures deemed to be offensive by social media users on platforms like Youtube and Pinterest and inspecting posts that have been flagged by users as hate speech or conspiracy theories, with final decisions often coming down to workers’ judgement calls. Search engine optimization done by digital workers for businesses impacts the results we see on our Google searches. Image classification work underpins the data sets used to train large machine learning models. The transnational flows of the platform economy, where work requests generated in the Global North are opaquely executed in the Global South, are reshaping the production of content and operational dynamics of the global internet. The phrase “microwork” tends to obscure the impact of such labour. The contractors hired to label data make decisions with far-reaching implications on what we see and what is kept hidden from our screens, yet there remains a lack of transparency on how these crucial decisions are being made.

Although platform-based labor has allowed African digital workers to access new sources of income, a growing pool of digital workers has led to downward pressure on wages and increasingly precarious working conditions. In competition with low bids from workers in Asia, particularly in India and the Philippines, African digital workers are forced to advertise their work for minimal wages and often reinvest their slim earnings in better internet connectivity, as internet data is still significantly slower and more expensive in Africa than elsewhere. Workers are at the whim of the platforms and their clients, who can cancel and reassign projects without prior notice and unexpectedly change payment methods and withhold wages. Platforms exert control over workers through rating systems and algorithmic workplace monitoring, with screenshots being taken of worker’s screens every 10 minutes to ensure they are glued to the task at hand.  

Discrimination toward African laborers is widespread on these platforms. Digital workers in South Africa, Kenya, and Nigeria report being confronted with clients who assume they are uneducated, unable to speak international languages, and willing to work for pay well below the national average. African workers will often mask or change their geographic location to attract the better job offers usually grabbed by European or North American workers. (A significant amount of platform work is also carried out in the United States, Canada, and the United Kingdom by people in dire economic situations looking for alternative routes to employment or extra money to supplement their social security checks.)

Regulating platform labor

There has been growing international attention toward the proper regulation of exploitative digital platform labor. The “Fairwork” index, a product of Oxford’s Internet Institute, ranks platforms according to factors that include fairness in pay, health and safety provisions, contracting, management, and representation. In January 2019, the International Labour Organization’s (ILO) Global Commission on the Future of Work called for the development “of an international governance system for digital labor platforms that sets and requires platforms (and their clients) to respect certain minimum rights and protections.” Part of the ILO criteria includes informing workers of their customers’ identity and the purpose of the work, with demands that platform operators mark tasks that may be psychologically stressful and damaging, such as content moderation work that exposes workers to graphic and violent images and has been linked to cases of post-traumatic stress disorder.  

There is currently limited to no government monitoring of the work performed on these platforms, despite their rapid growth. According to the ILO, approximately 10,000 new tasks are published and 7,500 are completed per hour on MTurk. By typically classifying their workers as self-employed, both clients and platforms can eschew responsibility for labor protections, social security benefits, and taxation. Although some EU countries, notably Denmark, are already in the advanced stages of rolling out systems to directly obtain income data from platform companies, North American regulators are lagging behind in their efforts to update labor rules to address this type of platform work.

Far from the promised prosperity and gilded dreams of a new digital economy, many African digital laborers are left facing economic precarity and psychological duress. The lion’s share of the value created by these platforms is carried back to corporate headquarters in Europe and North America, leaving digital workers behind. African states and global investors alike must protect workers from exploitative conditions if they desire to see African entrepreneurs prosper.

Beyond a moral imperative to uphold basic labor rights and prevent the growth of a new digital global underclass, U.S. government officials should be zeroing in on the type of work being performed by ghost workers, especially with recent mounting public pressure to combat online disinformation. Working in the shadows of the digital economy, these so-called “ghost workers” have immense responsibility as the arbitors of online content. The results of our Google searches, the tweets on our feed, and the product reviews we read are all the result this unseen labor. It is high time we regulate and properly compensate these workers.

Alexandrine Royer is the educational program manager at the Montreal AI Ethics Institute, a non-profit organization dedicated to democratizing AI ethics literacy.  

Amazon and Google provide financial support to the Brookings Institution, a nonprofit organization devoted to rigorous, independent, in-depth public policy research.