By Gituku Ngene

This article unpacks the issues raised in the Time Magazine article ‘Inside Facebook’s African Sweatshop’.

Earlier this year, Time Magazine investigated the operations of Meta’s partner, Sama, once again putting the spotlight on Africa’s nascent but rapidly growing microwork industry. Banyan Global defines digital microwork as a form of labour that relies on the internet infrastructure to break up complex technical projects into thousands of fractional tasks which can be completed within minutes. The tasks are completed by online workers, lending itself well to a global, remote labour pool. Micro workers are paid a nominal fee for each completed job, making the volume and speed of task completion an integral part of the microwork experience.

The industry has seen a steady rise fuelled by the growth of artificial intelligence (AI) and machine learning. The global AI industry alone is expected to grow by 26.1%, while the data annotation market is projected to reach over USD 13.5 billion by 2030. With AI, machine learning, and data annotation all on the rise, the global labour opportunity for digital microwork will grow, driving the offshoring of microwork to African countries. 

This investigation brought grave concerns about Sama’s work in Kenya and the sector in general. It led to a hot debate within the Jobtech Alliance community. Some argued that the article raised fair concerns about the decency and sustainability of content moderation work. Others highlighted that Sama’s reported low pay was a controversial, debatable subject, while others suggested that more is required to define regulations and standards to address the welfare of platform workers. These themes are not unique to Sama or Africa and point to universal issues in the digital work ecosystem. Last year, Meta and TikTok’s parent company – ByteDance, were sued by content moderators in class-action suits for failure to safeguard them from psychological risks posed by content moderation work.

The Time Magazine article therefore touched on a number of important issues. But it also left the reader with the inference that ‘microwork is bad’ and that it shouldn’t be promoted on the continent. This interpretation risks undermining the potential of a growing sector to create work for young Africans, and scaring potential service buyers from buying African talent. Therefore, we believe that it is critical to unpack the issues of the article separately. Unpacking these issues will allow them to be effectively considered and addressed independently, rather than conflating multiple issues into a blanket condemnation of an emerging sector. This article will unpack three key issues that stemmed from the article: 

1. Decency and dignity of work
2. Worker representation and unionisation
3. Fair pay across borders

Decency and dignity of work

In its investigation, Time revealed how Sama employees had been exposed to numerous hours of traumatic content moderation work, often viewing graphic videos for hours on end. This put them at high risk of contracting mental health disorders, compounded by the fact that established mechanisms to provide workers with psychosocial support and wellness breaks were reportedly ignored by managers in pursuit of high performance and exceptional delivery. Further, workers pointed out that they had been poorly informed about the nature of the jobs when applying. The firm is said to have advertised the positions as call centre agents and content moderators without disclosing the actual nature of the work to applicants. 

It is worth pointing out that most microwork is not sensitive content moderation. The ‘world of microwork’ covers everything from data entry, transcription, image tagging, virtual assistance, customer service and administrative support, to digital marketing, sales, and data analysis. Therefore a blanket tarnishing the microwork sector risks throwing the baby out with the bathwater. Nevertheless, the paradox of microwork is that while it allows for greater flexibility and convenience of work on the part of the worker, the breakdown of large jobs into micro-tasks results in work becoming very menial, banal and, in some instances, harmful. 

In this light, platforms need to re-think the deployment of this work and explore less strenuous models on the part of workers. One of the Jobtech Alliance members observes that such jobs could be made less miserable by mixing them up with other less strenuous tasks like image tagging, annotation, and transcription to ensure that these jobs only constitute a fraction of their work. This addresses the level of psychological strain for workers, but it could also allow workers to expand their skillset and set them on a career progression path to more complex, higher-paying jobs. Another Jobtech Alliance member questioned if this work could ever be ‘decent’, and if it was really suitable for any humans to do. Despite Meta pointing to artificial intelligence as the ultimate solution for such tasks, the reality is that the industry will require human intervention for the foreseeable future.

One of the Fairwork Foundation’s five principles on fair work centres on suitable conditions as a prerequisite for digital work to be considered fair and decent. Platforms should demonstrate that they have policies that:

1. Identify and address job-related health and safety risks (including psychological risks)
2. Flag any such risks to workers before they accept the job (e.g., indicating that potential exposure to violent content)
3. Provide clear reporting channels and outline penalties for clients who put workers’ health and safety at risk. 

In the case of content moderation, these risks are very high. Following these principle(s) would ensure that platforms clearly spell these risks at the point of hiring workers and, more importantly, provide meaningful support to workers, including significant psychosocial support.

Worker representation and unionization

Sama was accused of bashing workers’ efforts to push for better pay and work conditions by victimizing individuals who led and coordinated these efforts. Some workers were fired because of trying to unionize, while others were supposedly blackmailed into reneging their unionization plans. Some workers have already filed a lawsuit against Sama and Meta.

It is worth emphasizing that Sama’s business and operational model is not like that of most jobtech platforms – instead housed in an agency-style structure where workers still work within a centralized set-up. The complaints about anti-unionisation, if true, should be critiqued in the same way as the efforts of Amazon in the US or any other anti-union efforts.

But the theme does raise some interesting topics for the wider jobtech sector. One primary characterization of jobtech is the shift from traditional structures where the law provides for worker representation. In the absence of explicit legal provisions to govern the relationship between digital platforms and their workers, this has remained an increasingly complex legal grey area for platforms to navigate. The power dynamics of the platform–worker relationships often favoured the platforms and mediators of work, which has led to an increased push from workers to get mechanisms to channel issues like unfair pricing, long working hours and poor working conditions that platforms have perpetuated.

Africa’s microwork and freelancing industries are still playing catch up to their sister industries (taxi-hailing, gig platforms, e-commerce) on worker representation. Inevitably, we will likely see instances of more micro-workers pushing for more robust representation. Earlier developments have pointed to workers establishing informal outfits to champion their rights and grievances. Gradually these platforms formalize – growing from nondescript informal mobilization channels to fully-fledged labour unions. For instance, what started as ride-hailing convenings on WhatsApp groups in Kenya to organize strikes and push Uber to change conditions is now a legally established union – Kenya Digital Taxi Association which was behind a significant case against the platform(s) last year. Similarly, worker-led alliances have emerged to represent workers in South Africa, the UK and California. In Kenya, online workers have established the Online Professional Workers Association of Kenya. While still in their founding days, such entities are gaining popularity amongst digital workers and promise to shape the platform-worker relationship. 

Rather than fight unionization, platforms should see unions as key stakeholders in understanding and safeguarding the welfare of their workers. Platforms could collaborate in working with governments to build more explicit legal parameters for an industry that is still very nascent and undefined. Moreover, the jobtech sector itself offers opportunities to build alternative unionisation platforms as has emerged in Latin America and elsewhere. Jobtech could borrow lessons from the fintech industry, which has successfully structured regulation alongside government and other actors over the years, with fintech stakeholders playing a preeminent role.

The complexity of worker compensation across borders

The Time Magazine article raised concerns about the level of pay at Sama. At the time of the investigation, Sama paid its workers between $1.46 – $2.20 per hour, translating to a net monthly income of $440 – ($528). The article argues that Sama employees had expressed dissatisfaction with the pay – fuelled by the feeling that this compensation was not commensurate to the level of effort and strain that came with the work. Sama, pointed out that this pay was almost three times the minimum wage in the country. Past interviews of the late founder, Laila Jama, indicated that the entity regulates its wages in a way not to distort local labour markets as paying substantially higher would “throw everything off,” a position that’s controversial amongst industry pundits and critics.

One of the primary arguments for businesses in developed markets to outsource work to offshore markets is that such markets offer cheaper labour than developed countries, making them more attractive. However, the pay disparity between developed and developing countries stands out. While workers in Kenya earn as low as $1.46 per hour, their counterparts in North America or Europe would be earning over $10 per hour. However, pundits have pointed out that it is still three times the minimum wage in Kenya, and that most workers that take up such jobs come from Nairobi’s informal settlements where household incomes are generally significantly lower.

The issue of pay across borders is an incredibly complex one, and not one that we can answer in this article. Given that so many jobs can now be done anywhere in the world, do the justifications for differentiating by the cost of living still apply? In a recent article, Justin Norman of The Flip presents hugely varying positions on this in the industry, even within the US, with AirBnB paying its engineers the same wherever they are, and GitLab sticking firmly to its belief that it should pay people based on the cost of living where they are in the US, with the latter arguing that it enables them to hire more people. Similarly, without cost advantages, would these jobs – particularly for low-skilled roles – ever come to Africa in the first place?

One way that the local industry could leverage the current opportunities is to work towards graduating the sector into a higher class of automatically well-paying jobs. India is the poster child for this model – having started its offshoring industry in the mid-80s and growing it to a $150 Billion industry to date. By investing in skilling its labour force and building the right offshoring conditions, the country now attracts higher-skill jobs that translate to much-improved incomes for its workers. Similarly, China which started years ago positioning itself as a cheap labour destination has gradually matured its manufacturing industry to focus on higher-paying skilled jobs. The country achieved this by building its skills base and positioning itself more competitively in the manufacturing sector over the years. Is this a necessary process in the growth of the more inclusive tech ecosystem than just high skilled engineers? Could developing markets providing digital labour position themselves in a similar way?

All we aim to do here is present that this question of ‘fair rates for digital work’ is going to be an ever-present topic in the emergence of the sector, and explain the complexities in trading off the scale and the appeal of the market globally comparable rates. Unquestionably, platforms need to ensure that compensation is commensurate with the nature of work and skillsets of the workers – and accounts for factors such as extra shifts and demanding targets because of electronic monitoring. Interestingly, two weeks after the article was published, Sama announced that it had increased employee salaries by 30% to 50 to $2.20 per hour (approx. $439 monthly net).

Moving forward

The microwork industry ecosystem in Africa is still in its very early days but holds high potential to create thousands (if not millions) of jobs for young Africans. In such a nascent ecosystem, there will be a number of questions that need to be answered. Articles such as the Time Magazine one brought useful attention to some of the sensitivities of this emerging sector (and resulted in pay increases for Sama workers!), but it is critical to understand and respond to the different issues in a coherent way to avoid bundling them into one broad negative picture of the industry- otherwise, we risk throwing the baby out with the bathwater.

The author is the Senior Advisor for Youth Employment and Innovation at Mercy Corps. All views expressed are that of the author.