On AI colourisation: algorithms, ancestry, and colour beyond the black box

This article investigates the recent fusion of AI colourisation with genealogy and ancestry databases to offer a set of reflections on this nascent technology. Combining humanistic approaches from film and media studies, science and technology studies, critical race studies, and visual and material culture, it attempts to disentangle the political, technical, and aesthetic concerns that arise when the achromatic past becomes the colourised present through machine vision. After tracing the computational origins of colourisation, the article reveals how deep learning-based colourisation tools mark a rupture in the way the machine 'senses' colour, where the logic of pattern recognition and classification overrides epistemologies of sensory perception. The final part of the essay turns to the racialised role colourisation occupies on genealogy platforms, arguing that such databases naturalise the historically fraught relationship between colour as both race and hue. Across these three sections, colour is at the centre of these questions of subjectivity, personhood, and technologically mediated ways of seeing. It remains the vexed and ambivalent site to which meaning adheres.