Rights groups are calling on Meta Platforms to seize the opportunity to improve its content moderation in Africa after its main third-party contractor in the region said it would no longer screen harmful posts for the social media giant.
Kenya-based outsourcing firm Sama said on Jan. 10 it would no longer provide content moderation services for the owner of Facebook, WhatsApp and Instagram in March as it moves to concentrate on data labelling work.
Sama said it would be laying off 3% of its staff – about 200 employees – to streamline its operations and boost efficiency. It will continue to provide data labelling services to Meta.
The announcement comes as both Sama and Meta face a lawsuit over alleged labour abuses and preventing workers from unionising in Kenya. Another lawsuit accuses Meta of allowing violent posts to flourish on Facebook, inflaming civil conflict in neighbouring Ethiopia. Both companies have defended their record.
Digital rights campaigners said efforts made by Meta to curb harmful content in African countries were woefully inadequate compared to richer nations, and called on the company to drastically improve its moderation processes.
“With the exit of Sama, now would be a good chance for Meta to put things right and ensure better labour conditions for African moderators in the region,” said Bridget Andere, Africa policy analyst at Access Now.
“Meta should increase the number of moderators for the region to adequately cover local languages and dialects, and also be more transparent about their algorithms which are promoting harmful content,” she told the Thomson Reuters Foundation.
Meta did not provide any details on whether it had found a new third-party contractor for East Africa, but said Sama’s withdrawal would not adversely impact users on its social media platforms.
“We respect Sama’s decision to exit the content review services it provides to social media platforms,” a Meta spokesperson said.
“We’ll work with our partners during this transition to ensure there’s no impact on our ability to review content.”
Meta’s legal challenges in East Africa began in May last year after former moderator Daniel Motaung filed a lawsuit over poor working conditions in Kenya.
The petition, also filed against Sama, alleges that workers moderating Facebook posts faced irregular pay, inadequate mental health support, anti-union activity, and violations of their privacy and dignity.
Sama denies the allegations, while Meta said it requires “its partners to provide industry-leading pay, benefits and support to workers”.
A judge is expected to rule on Feb. 6 on whether a Kenyan court can hear the complaint.
Last month, Meta was landed with another lawsuit that accuses the company of allowing violent posts to flourish on Facebook, inflaming Ethiopia’s civil war.
The lawsuit, filed by two Ethiopian researchers and Kenya’s Katiba Institute rights group, argues that Facebook’s recommendations systems amplified hateful and violent posts in Ethiopia, including several that preceded the murder of the father of one of the researchers.
The plaintiffs are demanding that Meta take emergency steps to demote violent content, increase moderation staffing in Nairobi, and create restitution funds of about $2 billion for global victims of violence incited on Facebook.
Meta said it has strict rules outlining what is and is not allowed on Facebook and Instagram.
“Hate speech and incitement to violence are against these rules and we invest heavily in teams and technology to help us find and remove this content,” the Meta spokesperson said.
“Our safety and integrity work in Ethiopia is guided by feedback from local civil society organizations and international institutions.”
END OUTSOURCING, REVIEW ALGORITHMS
Globally, thousands of moderators review social media posts that could contain violence, nudity, racism or other offensive content. Many work for third-party contractors rather than directly for tech companies.
Meta has faced previous scrutiny over content moderators’ working conditions, and criticism of its action to stop hate speech and violent content.
In July 2021, a California judge approved an $85 million settlement between Facebook and more than 10,000 moderators who accused the firm of failing to protect them from psychological injuries due to their exposure to graphic and violent imagery.
The same year, Rohingya refugees from Myanmar filed a lawsuit, suing Meta for $150 billion over allegations the company did not take action against anti-Rohingya hate speech which contributed to violence against the minority group.
Rights groups say moderation processes fail to understand specific cultural and societal contexts, and lack knowledge of local languages and dialects, allowing problematic content to spread quickly and be amplified with serious consequences.
Cori Crider, director of Foxglove – a London-based tech-justice advocacy group which is supporting both lawsuits in Kenya – called on Meta to end its outsourcing practices and increase the “pitiful” resources it puts into safety in Africa.
“Facebook should take moderation in-house, hire every one of the 260 content moderators in Nairobi who do vital safety work, and value these people with decent wages, full clinical mental health support and dignity and prestige,” Crider said.
“Outsourcing is broken and both workers and ordinary people have paid the price.”
Alia Al Ghussain, AI and human rights researcher at Amnesty International, said Facebook’s algorithms – which power its news feed, ranking, recommendations and groups features, shaping what users see on the platform – also needed scrutiny.
“Meta’s engagement-based business model means algorithms are programmed to cause content which is harmful and inflammatory to go viral,” said Al Ghussain, whose organisation is supporting the lawsuit over Ethiopia.
“The litigation that we are seeing both in the Ethiopia case and the Myanmar case clearly demonstrate that the problem is systemic and urgently needs to be addressed through regulation.”