A Real Monet Suspected of Being AI

On May 12, 2026, the artist SHL0MS posted an image of a Monet painting on X. He claimed to have generated it with AI in Monet’s style and asked viewers to explain, in as much detail as possible, why it was inferior to a real Monet. Replies criticised its composition, colours and brushwork, among other features. The commenters were actually looking at a reproduction of the painter’s work. The German newspaper FAZ discussed the deception in a commentary, questioning its value as media criticism. [1][2]

The underlying effect has been studied scientifically. A 2023 paper in Scientific Reports, comprising six experiments with a total of 2,965 participants, found that simply labelling art as AI-made can influence its evaluation. Even when the images were identical, attributing them to AI led to lower ratings, particularly for artistic skill and monetary value. [3]

The Monet stunt itself is not a controlled experiment. Its leading question explicitly invites negative judgements, and the publicly visible responses are not a representative sample. Nor does it establish that criticism of generative AI is unfounded.

It nevertheless offers a revealing sign of the post-generative present: the possibility of machine-generated images is now changing how historical art is viewed. Judgement encompasses both the image and the presumed process behind its creation.