Practice or Product? Labour, Training, and the “Ethics” of AI

This essay interrogates the risks posed by artificial intelligence (AI) to intellectual labour and human skills and capacities. Employing insights from ethical philosophy, political theory, and the Marxist tradition in an engagement with recent manifestos calling for AI regulation, I adopt an interdisciplinary approach to the question of labour and technology, using translation work as a case study. Intentionally bracketing whether AI will be able to replace human translators, I explore the foundation of the conflict between AI and human intellectual work: namely that the former risks destroying the cultural practices and institutions that maintain the human ability to think and communicate in the most general sense. Even if regulation were to succeed in making AI more “ethical” – that is, more transparent, less exploitative, less biased, and less environmentally destructive – it would still be “unethical” in the strict etymological sense of the term that I advance here as a hermeneutic device: AI destroys the ethos (habits, abilities, way of being) of translation by degrading the cultural and institutional “training milieu” conducive to it. This conclusion is applicable to numerous domains of labour and has implications for education and democratic citizenship.