Rethinking AI hermeneutics: from interpretative agents to projected worldviews

An open discussion in digital hermeneutics is whether interpretation remains an exclusively human activity in a digital world. Positions range from viewing interpretation as an existential human practice to considering digital technologies' role in shaping interpretation. Some further open up to recognising the interpretative agency of digital technologies, especially when it comes to artificial intelligence (AI) systems. While we agree AI systems frame the world and act in it, as they classify, rank, correlate, and structure how things appear, we rather argue that they do it by repeating and amplifying patterns inherited from large corpora of historical data and under – humanly – predefined objectives; not as novel interpretations of the world. In this chapter, we thus reframe this discussion in terms of design responsibility. We retrace the specific way AI systems frame and pre-structure the world and afford meaning-making, and underscore how this results from design decisions, supported by political and economic aspirations. To show the eco-socio-material implications of AI systems, we discuss how they are grounded in colonial and extractive logics and categories, outsourced digital labour, and hegemonic knowledge production. We conclude by showing how AI narrative and hype conceal the political project of AI and its perpetrators, and that framing AI as an autonomous interpretative agent – a form of non-anthropocentric hermeneutics – risks doing the same. As we've reframed this discussion as a design responsibility one, we find a glimpse of hope in design processes centred around participation and social justice.