An Artificial Intelligence analysis of two paintings by the world-famous Flemish artist Jan van Eyck has raised an important question. Are the paintings really his work? Experts say they may be studio works that originated in Van Eyck's studio but were not painted by him. 'That could mean that there is a lost original somewhere that should be seen as more Van Eyck's work than these examples. The two paintings in question were once owned by the Bruges nobleman Anselm Adornes.
“Saint Francis Receiving the Stigmata” is the title of two almost identical, unsigned paintings attributed to Jan van Eyck, one of the greatest painters the Low Countries has ever produced. Worldwide, he is renowned for his keen eye for detail, his use of light and his oil painting technique. Only about 20 works from his oeuvre have survived.
Stigmata refers to the spontaneous appearance of bodily wounds, scars, or pain in locations corresponding to the crucifixion wounds of Jesus Christ. Primarily associated with Roman Catholicism, these marks are considered a mystical phenomenon, representing a physical union with the suffering of Christ.
One version of “Saint Francis Receiving the Stigmata” hangs in the Musei Reali in Turin, Italy, the other in the Philadelphia Museum of Art in the United States. Art lovers flock to admire them as happened when the works were on display at the major Van Eyck exhibition at the MSK Fine Arts Museum in Ghent in 2020.
But what if neither of these works is really a Van Eyck? That is a question that has arisen following a scientific analysis using artificial intelligence (AI) conducted by the Swiss company Art Recognition.
Art Recognition uses AI to verify the authenticity of paintings and drawings. The company works for auction houses, collectors, galleries and art dealers, and also carries out research projects, for seats of learning like Tilburg University in the Netherlands.
Art Recognition compares digital photographs of a work with a huge database of known, authentic works by the artist. In the case of “Saint Francis Receiving the Stigmata”, Art Recognition says that the version in Philadelphia was “91 per cent negative” and the version in Turin was “86 per cent negative”.
This would suggest that there is little certainty that Van Eyck actually painted these works himself.
So are they studio paintings?
A few weeks ago the analysis of Van Eyck's work was discussed on the podcast “Is it? The Art Mystery Podcast” with American art historian Noah Charney. He is also an author and is regularly asked to act as an expert in cases of suspected art forgery or crime.
In each episode, he asks whether a specific painting was really created by the famous master. Charney was surprised when he saw the analysis of “Saint Francis Receiving the Stigmata”. “The negative result for both works was so astonishing that in-depth tests were carried out to confirm the results,” he says.
Charney spoke with Till-Holger Borchert, former director of Bruges Museums Musea Brugge and one of the foremost Van Eyck experts. According to Borchert, the analysis supports earlier suggestions by experts that both versions of “Saint Francis Receiving the Stigmata” are studio paintings. This means that they were made in the artist's studio, but not necessarily by the artist himself.
Borchert also felt the high negative score was pretty surprising but says that the AI analysis of the artworks mainly raises questions that need to be investigated further.
Is there a lost original out there somewhere?
Charney himself thought that the painting in Turin would be confirmed as an original, and the version in Philadelphia as a copy, originating from Van Eyck's studio or from a later period. ‘The negative results suggest that the two versions are studio works,’ says Charney.
‘If a work comes from Van Eyck's studio, that does not necessarily mean that he actually painted all aspects of it himself,’ he explains. ‘That's a misconception that people have retained from the 19th-century image of the lonely artist in an attic room in Paris drinking absinthe, smoking cigarettes, wearing a beret and doing every aspect of the work himself.’
‘All this could mean that there is a lost original somewhere that is much more Van Eyck's work than these copies.’
Once belonged to Bruges aristocrat Anselm Adornes
Both Francis paintings also have an additional link with Flanders. In 1470 the Bruges nobleman and businessman Anselm Adornes mentions the works in his will. He was about to embark on a pilgrimage to the Holy Land, feared for his life, and bequeathed a painting of Saint Francis by Van Eyck to each of his two daughters. His daughters Margriet and Louise were both nuns; it is not clear whether they ever received the paintings. Anselm Adornes returned safely from his pilgrimage, as evidenced by the Jerusalem Chapel and the Adornes estate in Bruges.
Rubens's “Samson and Delilah” and fake masterpieces on eBay
Art Recognition has built up quite a reputation since it was established in 2018. In 2021, with the help of AI, the company concluded that Peter Paul Rubens's masterpiece “Samson and Delilah” was 91 per cent likely not to have been painted by the Flemish baroque master. There had long been doubts about the work, which is on display at the National Gallery in London.
In 2024, Art Recognition was able to identify dozens of suspected fakes that were being touted as authentic on the auction site eBay. These included works by Claude Monet and Pierre-Auguste Renoir. The Monet was offered for $600,000, the Renoir for $165,000, but were 95 per cent likely to be fake.
This was a boost for Carina Popovici, co-founder and director of Art Recognition, who wants to use AI analyses to detect forgeries in the art world with greater accuracy and protect buyers. At the same time, AI should help museums, collectors and auction houses to make the verification process of artworks more objective, accurate and faster.
The high negative results for “Saint Francis Receiving the Stigmata” also surprised her. She had thought that if one painting was negative, the other would be positive. ‘But no, they were both negative,’ she told UK daily The Guardian.
‘I don't think the museums in Philadelphia and Turin will be happy with this. It's not good news for them.’
The Guardian contacted both museums for comment but has not yet received a response.
Art Recognition previously conducted an analysis of “The Portrait of Giovanni Arnolfini and his Wife”, another famous work by Jan van Eyck. That analysis yielded an 89 per cent probability that it was authentic.