Facebook is facing yet more headaches over its content moderation in Kenya after human rights group Global Witness revealed that the social media giant had failed to detect hate-speech ads in the run-up to the general election on 9 August.

Facebook approved 20 adverts that included hate speech and incitement, which Global Witness had submitted as test cases. Meta, Facebook's owner, boasted only last month about its readiness for the elections, waxing lyrical about how it was investing in people and technology.

The damning findings elicited a swift response from the National Cohesion and Integration Commission (NCIC), which threatened to recommend a ban on Facebook. The government immediately stepped in to reassure Kenyans - and the world - that it would not ban the platform or shut down the internet during the election.

Social media and internet shutdowns are common during African elections - a move that, thankfully, Kenya has so far refrained from. The NCIC's threat over Facebook's failures was overzealous and worrying, but the issues that motivated it are alive and urgent.

The NCIC said that Facebook's inaction in curbing the spread of "hate speech, incitement, misinformation and disinformation" is a violation of Kenya's constitution. It reportedly demanded that the company increase the number of content moderators in the country, especially those who can assess content in Indigenous languages.

But Facebook is already facing legal action over its content moderation - for its allegedly problematic labour practices.

Normalising exploitation in Africa

In order to keep vile content off its platforms' feeds, Meta has thousands of subcontracted content moderators in the US and outsourcing hubs such as Manila in the Philippines and Nairobi in Kenya. They're working for low wages, under stressful conditions, with lingering traumatic effects. This has resulted in numerous lawsuits being filed against the company, in the US, Europe and, most recently, Kenya.

Daniel Motaung, a former Facebook content moderator hired by Sama - an outsourcing company that dubs itself an "ethical AI" outfit - filed a lawsuit against the two companies in Kenya in May. The lawsuit alleges that, in addition to paying content moderators far less than elsewhere for the same work, Sama also engaged in forced labour, human trafficking and union busting.

Sama has denied the allegations, which a spokesperson said are "disappointing and inaccurate". Speaking to British newspaper Metro, they added that the company offers "competitive wage, benefits, upward mobility, and a robust mental health and wellness program".

After the lawsuit was filed, a Meta spokesperson told Reuters: "We take our responsibility to the people who review content for Meta seriously and require our partners to provide industry-leading pay, benefits and support. We also encourage content reviewers to raise issues when they become aware of them and regularly conduct independent audits to ensure our partners are meeting the high standards we expect."

Sama's Africa operations are headquartered in Nairobi. When Facebook announced in 2019 that it was setting up content review operations in the Kenyan capital via Sama (then Samasource), I cautioned against celebrating this as a 'win' that was bringing digital jobs to the continent. There was more to it than met the eye, I felt.

My misgivings appeared to be validated when Motaung's account of his time as a content moderator in Nairobi was published earlier this year. It was a debilitating story of how adverse labour practices around content moderation have made their way to Africa. Since then, more moderators in Kenya have asked that non-disclosure agreements that hinder them from testifying in court be lifted.

Samasource's founder, Leila Janah, who passed away in January 2020, once defended paying low wages to African workers by saying the company did not want to distort local labour markets. Importing problematic work and replicating the local market's own problems (such as low wages) is, I guess, the Big Tech way.

During the pandemic, this kind of work was even glamourised, with Sama being profiled in Wired magazine for moving workers to hotels so that they could continue doing data labelling for firms such as Google, despite the lockdown imposed by the Kenyan government.

As the Wired article says, Sama has branded its contracts to recruit low-wage workers in Africa for these kinds of 'janitorial' tasks for Big Tech, "a kind of aid program", with which it is lifting up people in the poorest places on earth.

Following Motaung's whistleblowing, Sama has offered a pay rise to its content moderators in Kenya. Meanwhile, Meta's lawyers offered the standard Big Tech defence in developing countries - that the company cannot be sued in the country because it is "not resident, trading or domiciled in Kenya".

In 2020, without admitting wrongdoing, Meta agreed 'in principle' to pay $52m in compensation to more than 11,000 content moderators in the US who had developed mental health issues on the job. This content moderation side-effect is being exported to Africa.

Motaung's lawsuit is seminal because it's one of the first outside the West to be brought against the social media giant. The court's decision will set an important precedent for how Meta and its peers are held accountable for their labour practices, even when outsourced. It is encouraging that Kenyan courts have already issued a landmark ruling against another tech giant, Uber.

It should also be a wake-up call to African policymakers to set guardrails for how the gig economy is regulated on the continent.

Big Tech-adjacent outsourcing firms such as Sama should not be allowed to perpetuate problematic labour practices, while also making a pretty penny and amassing accolades for 'doing good' as they bill themselves as social enterprises helping the developing world.

Africa may need jobs and have a digitally savvy workforce ready to take on opportunities, but we must reject normalising these exploitations.

From openDemocracy