Source: The Next Web
Meta terminated its contract with Sama, a Kenyan data labeling firm, after Swedish journalists revealed that low-wage workers were reviewing sensitive video feeds from Meta's Ray-Ban smart glasses. The move exposes the company's reliance on human annotation for AI training and the vulnerability of workers who see proprietary product data before public release. The termination amounts to damage control rather than an ethical reckoning. Meta's public commitment to responsible AI development depends on a precarious chain of contractors in the Global South who lack legal protections and can be dismissed once their labor becomes a liability. The pattern runs deeper than one vendor relationship. As consumer AI products embed cameras and sensors into everyday objects, the human work required to manage the gap between technical capability and acceptable privacy standards moves offshore and out of sight.