Seeing is making: AI visualisation and genomic prediction

The integration of artificial intelligence (AI) into genomics is reshaping not only how biological data are analysed, but how genomic knowledge is produced and operationalised in clinical practice. Earlier computational approaches relied on alphanumeric outputs—risk scores, statistical associations and textual reports—that required interpretive reasoning to translate data into clinical meaning. By contrast, contemporary AI systems increasingly generate visual outputs such as maps, rankings and image-based representations that render genomic information immediately perceptible as clinically relevant futures.

This paper argues that this shift from alphanumeric processing to visual forms reconfigures the role of interpretation in genomic reasoning. Instead of requiring clinicians to reconstruct the inferential steps linking data to conclusion, AI systems present structured visualisations that foreground outcomes as ready for action. In this context, visualisation does not simply display results but participates in organising what counts as knowledge in the first place. As a result, genomic modelling no longer functions primarily as a predictive framework grounded in explainable evidence, but as a system that presents actionable futures whose authority lies in their visual form, raising the question of how clinical action is being grounded when these images shape the very biological outcomes they appear to represent.