Meta terminates contract with Sama, triggering 1,108 redundancies in Kenya over smart glasses content dispute
Workers allege they were forced to review graphic footage from Meta's AI glasses, prompting investigations by regulators in the UK and Kenya
Meta has ended its contract with data annotation firm Sama, a move that will result in the redundancy of over 1,100 workers based in Kenya. The decision follows a dispute between the two companies regarding the nature of the content reviewed by staff and the reasons for the contract's termination. While Meta attributes the end of the relationship to Sama's failure to meet operational, security, and quality standards, Sama firmly disputes this assessment.
In a statement released to the BBC, a Meta spokesperson confirmed that work with Sama was paused last month to investigate specific claims before the contract was ultimately terminated. Meta maintains that users provide clear consent for their photos and videos to be reviewed by humans to improve product performance. However, Sama asserts that it has consistently met all required standards across its engagements with Meta and claims it was never notified of any failures prior to the decision.
The controversy stems from allegations made by workers at Sama, who revealed in interviews with Swedish newspapers that they were required to manually label content captured by Meta's new smart glasses. These data annotators described reviewing graphic material, including footage of people having sex and instances of non-consensual recording, as part of their duties to train the artificial intelligence. A Kenyan workers' organisation has since alleged that Meta's decision to cut ties was a form of retaliation for staff speaking out about the nature of the content they were exposed to.
Regulatory bodies in both the United Kingdom and Kenya have launched inquiries into the privacy concerns raised by the smart glasses following the initial reports. The UK Information Commissioners Office wrote to Meta describing the reports as "concerning," while Kenya's Office of the Data Protection Commissioner announced it is commencing its own investigation into the privacy implications of the devices.
Sama, a US-headquartered outsourcing business that operates as an ethical B-corp, previously moderated Facebook posts in a deal that attracted significant criticism and legal action from former employees. Naftali Wambalo of the Africa Tech Workers Movement, who represents petitioners in related legal cases, suggested that the standards Meta cited were merely a pretext to silence workers discussing the reality of their labour conditions.
As the situation unfolds, the incident highlights the tensions surrounding the deployment of AI hardware and the ethical considerations of human oversight in data annotation. Lawyers representing the affected workers have warned that relying on such a "flimsy foundation" for the AI ecosystem poses significant risks for the Kenyan government and the broader industry.
