Earlier this year, Meta and its largest content moderation partner in Africa, Sama, were accused of union busting, forced labor, and human trafficking. The lawsuit claims that “misleading job ads” lured potential employees from across Africa who, once realizing the true nature of the work, often had no means to get home. And when content moderator Daniel Motaung attempted to organize his colleagues for better working conditions and pay, Sama fired him.
A win for Motaung, who filed the lawsuit, could force social media companies to invest in their content moderation workers, even if they’re not direct employees. (In response to the lawsuit, Meta claims they never employed Motaung and are therefore “not liable for or privy to” any of the allegations. However, Motaung argues the moderators are Meta employees in a material and legal sense: they use Meta’s internal systems and guidelines, work closely with Meta staff and on a schedule of work set by Meta.) What hasn’t received as much attention, though, is what the lawsuit means for enterprises claiming to improve the developing world. Sama is a so-called social enterprise founded specifically to offer “decent work” to low-income people globally. Definitions of “social enterprise” vary, but most academics and entrepreneurs agree that they aim to maximize revenues and profits while contributing to a social or environmental goal—usually, by supporting a specific marginalized group. In Sama’s case, this is their employees, who often have little to no prior experience in the formal economy. A self-proclaimed “ethical AI” company, Sama has been lauded by Fast Company, B Corp, and Forbes, among others. The fact that Sama is now accused of abusing the very workers it tried to empower reveals the fundamental brokenness of the social enterprise model.
Legal context first: The lawsuit was lodged in Kenya, which has relatively weak labor protections that the government has often failed to enforce. Government workplace inspections remain rare, courts face significant backlogs, penalties tend to be incommensurate with the offense, and employers often fail to comply with court orders. For these reasons, it’s rare for employees to file complaints at all. Even if Motaung wins his lawsuit, prompting a new set of standards for content moderation work, there’s no telling if these standards will actually be implemented in Kenya.
Seen in that light, setting up a regional content moderation hub in a place with such weak labor protection seems almost strategic, or at least convenient, for Meta. Payroll savings aside, no Ministry of Labor official monitored what staff were actually moderating: usually highly disturbing content including beheadings and child sexual abuse, according to Motaung. Meta’s name didn’t even have to be on the door. As a contractor hired to moderate Meta’s content in Africa, it was Sama who recruited and technically employed the workers–approximately 240 in their Nairobi office. The company specializes in data annotation and digital microwork that can be performed by low-income people in the developing world. On top of content moderation, the company also offers image, video, and other product annotation services for clients including Google, Walmart and Getty Images.
Maybe Sama’s current problems began with a fundamental change in mission: Initially founded as the non-profit “SamaSource” in 2008, the company converted to a for-profit social enterprise structure in 2019. Making money became just as much, if not more, of a priority as providing decent work. Evidence of this internal mindshift can be seen in Sama’s documents: early SamaSource reports are filled with references to giving people “dignified” work and measuring impact in terms of changes in workers’ lives and communities. But fast forward to its transformation to a for-profit and its subsequent rebranding to “Sama,” and this focus on worker impact seems to have, if not disappeared, at least receded.
The company has always claimed to pay workers a “living wage,” which usually exceeds the minimum wage and ensures a decent standard of living for employees in a given country. During the early to mid 2010s, Sama workers in Kenya earned $8 a day, roughly in line with estimates of living wages for that time period. And a randomized control study found that Sama’s training and job referral program did have long-term benefits on workers’ employment and earnings, even after they left Sama. However, a recent TIME investigation found Sama’s lowest paid workers in Nairobi earned only $1.50 an hour–barely above Kenya’s current $1.15 minimum wage for cleaners, and well below the $2.61 an hour that cashiers must be paid. Finding “a workplace culture characterized by mental trauma, intimidation, and alleged suppression of the right to unionize,” with Sama workers numbering among the lowest-paid employees for Meta anywhere in the world, TIME’s investigation also calls the RCT findings into question.