A painting, attributed to Raphael by AI earlier this year, now faces doubts from other AI and art experts.
The painting in question, de Brécy Tondo, has long been contested for its authenticity. While some claimed that it was a lost work of the Italian maestro Raphael, others claim that it was a copy made during the Victorian era. Yet others claim that while the work might indeed have been created during the Renaissance, it is still a copy.
However, researchers at the University of Nottingham and the University of Bradford published their results earlier this year, seemingly ending the discussion. By having artificial intelligence analyze the painting and another Raphael work, the Sistine Madonna (made in 1513 and currently hanging in Dresden, Germany), they came to a striking conclusion. There is a 97 percent likelihood that the maestro painted de Brécy Tondo, the report said in January.
Since then, the painting has gone for display at the Cartwright Hall Art Gallery in Bradford, England, where it was unambiguously attributed to the Italian artist for the first time. The gallery also praised the research, saying “Artificial intelligence-assisted computer-based facial recognition showed the faces in the paintings are IDENTICAL to those in Raphael’s famous altarpiece.”
However, now there are dissensions regarding this conclusion. YetCarina Popovici, a scientist at the Swiss company Art Recognition, has said that her own findings, again with the aid of AI, were entirely dissimilar. She found an 85 percent likelihood that de Brécy Tondo was not painted by Raphael. On the other hand, Timothy Clifford (former director-general, the National Gallery of Scotland) has entirely dismissed the use of AI to analyse art. According to him, there was a “terrible unlikelihood that the AI was remotely correct”, and called it dangerous to use “mechanical means” for art attribution.