Visibility as value: The changing economics of translation

AI-powered linguistic technologies, particularly neural machine translation and generative AI, have profoundly disrupted the language services market since 2016. While demand for translation grows, it is increasingly embedded as a free, commodified feature on platforms, operated by a handful of very large tech companies. In parallel, financial value increasingly stems from data generated by user interactions and not linguistic content per se. This article examines the role of invisibility and opacity in the reconfiguration of economic value in the language industry, focusing on three interconnected dimensions: the digital economy, the shift from content to data, and labour conditions. The opacity of data collection and exploitation processes has allowed for large-scale appropriation of human content and raises issues of copyright. Meanwhile, the labour required to train and maintain these systems — for the most part low-skilled, crowdsourced, and poorly remunerated — remains largely invisible, which exacerbates inequalities and deskilling. The article argues for greater scrutiny of generative AI’s business models and calls on stakeholders to document these practices and to challenge the narrative of automated efficiency.