Realism In The Age of the Simulated Image

From Renaissance optics to the development of photography, and from bitmap images to deepfakes, the mutability and indexicality of images have undergone significant transformations in both popular imagination and artistic and scientific communities. In contemporary visual culture, the emergence of the 'deepfake'—a photorealistic image without a physical referent—has challenged traditional notions of photographs as indicators of truth or evidence. Although we are aware that digital images can be manipulated at the pixel level, the photograph has long persisted as a source of evidence until machine learning provided the ultimate alibi for ‘(deep)fake news.’ In the realm of visual studies and scientific imaging, the ability to alter indexical images to enhance communication has been a topic of intense debate and scrutiny. Does the addition of ‘false color’ to electron microscopy images diminish their realism, or does it enhance their communicative power by incorporating color data? Moreover, does the manipulation of such indexical images transform them into something resembling a diagram or illustration, or does the principle of indexicality and realism already allow for a certain degree of manipulation? This presentation will overview these debates and, through a technical comparison of two algorithmic images of black holes, argue that algorithmic proof-of-work is presently a more relevant form of 'realism' than other formal criteria. By emphasizing the implicit reliance on algorithmic realism in both scientific and popular images, this presentation contends that we have already transitioned to a standard of representation based on the manner in which an image was created rather than its appearance based on how an image was created, rather than what the image looks like.