Artificial Intelligence used in art attributions: AI analysis on Rubens painting suggests it is not by the Flemish master

The Swiss company Art Recognition has carried out AI analysis on Rubens’s Samson and Delilah at the National Gallery of London. The result suggests there is a 91.78 percent probability the painting is by another hand.

The National Gallery purchased Samson and Delilah at Christie’s in 1980 for the considerable sum of £2.5 million (equivalent to £6.6 million today). It quickly became one of the National Gallery’s ‘highlights’, and on their website they suggest the painting shows Rubens responding to his trip to Italy, capturing the “atmosphere of ancient Rome” and showing the influence of Caravaggio and his experimentations with chiaroscuro.

However, the attribution of this painting has long since been debated. It is known that Rubens did create a painting of this scene in 1609 or 1610 for his Antwerp patron, Nicolaas Rockox. A contemporary engraving after Rubens’s painting exists by Jacob Matham, and a work by Frans Francken the Younger documents Rubens’s painting: it shows a banquet at Rockox’s house, in which Rubens’s masterpiece is proudly displayed in the lavish interior. Following Rockox’s death in 1640, the painting disappeared. The National Gallery version emerged in 1929 with an attribution to Gerrit van Honthorst. Rubens expert Ludwig Burchard signed a certificate of authentication claiming the work was by Rubens. Yet critics noticed a number of problems with the attribution. Firstly, the National Gallery version differs from the engraving and Frans Francken the Younger’s painting, with Samson’s foot cropped. It has also been suggested that the palette is uncharacteristic of Rubens.

Euphrosyne Doxiadis, an independent scholar and artist, stated that: “when I first saw the National Gallery’s Samson and Delilah in 1987, immediately I thought it could not have been painted by Rubens and I supposed that it was a copy—a 20th century copy.” It is of further note that, following the death of Ludwig Burchard – the scholar who first made the attribution to Rubens – it emerged that he was falsely authenticating works for his own financial gain. Many of his Rubens attributions have since been amended. However, not all critics have agreed. Art historian and previous director of the National Gallery, Neil MacGregor, suggested that the painting, “is very different from his [Rubens’s] later works,  but not unlike his other paintings of that period… We stand by Samson and Delilah as an unusual but authentic Rubens.”

Dr. Carina Popovici, the scientist who carried out the AI analysis, said: “the results are quite astonishing […] we repeated the experiments to be really sure that we were not making a mistake and the result was always the same. Every patch, every single square, came out as fake, with more than 90% probability.” In comparison, the same AI analysis was carried out on another painting in the National Gallery’s collection by Rubens, A View of Het Steen in the Early Morning, 1636, which obtained a probability of 98.76 percent in favour of an attribution to Rubens.

This is not the first time artificial intelligence has been used in the art world. Recently, the Rijksmuseum has used AI to reconstruct parts of Rembrandt’s The Night Watch, and the technology has also been used in Picasso research.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s