An artificial intelligence company claims to have outsmarted decades of research by expert art historians using only a photograph.The Swiss AI research company Art Recognition says a painting long believed to be a copy of a work by Caravaggio is in fact an original by the Baroque Italian painter, disputing long-held attributions by Sotheby’s, the Metropolitan Museum of Art, and others. The company’s AI analysis of “The Lute Player,” formerly held in the Badminton Estate in Gloucestershire, concluded that there is an 85.7% likelihood that the work is a bona fide Caravaggio, the Guardian reported on Saturday, September 27. Art Recognition boasts on its website that its system can verify an artwork in a few days, “based only on a photograph of it.” According to a list of so-called case studies on its website, the firm claims to have authenticated works by Raphael, Vincent van Gogh, and Anthony Van Dyck in collaboration with academic institutions like Tilburg University and the University of Liverpool.“Everything over 80% is very high,” Carina Popovici, Art Recognition’s co-founder and chief executive officer, told the Guardian, commenting on the AI’s alleged accuracy rate.The claim is all the more eyebrow-raising because very few paintings by Caravaggio are believed to have survived. The artist, renowned for his masterful use of dramatic chiaroscuro and his dynamic renderings of violent scenes that mirrored his own notoriously volatile lifestyle, died in 1610 at age 38. “The Lute Player,” one of three versions of a painting depicting a brown-haired boy playing the instrument at a marble table adorned with flowers, was initially acquired in the 18th century by Henry Somerset, the Third Duke of Beaufort of the Badminton Estate. The other two paintings, both made around 1596, are located in the State Hermitage Museum in St. Petersburg and in the private Wildenstein Collection, headquartered in New York, respectively.State Hermitage Museum version of “The Lute Player” (image CC BY-SA 3.0 via Wikimedia Commons)The State Hermitage Museum version is an undisputed Caravaggio that also features a marble slab with floral displays. Meanwhile, the Wildenstein version, which has a decorative table setting topped with musical instruments, has been met with skepticism over its authenticity, namely from lute maker and Lute Society President David Van Edwards.Between 1990 and 2013, the Wildenstein version was on an extended loan to the Metropolitan Museum of Art, where then-curator of European Paintings Keith Christiansen declared its “authorship and provenance beyond any doubt” in an exhibition catalogue for the 1990 show A Caravaggio Rediscovered: The Lute Player. The museum traced it back to the collection of Cardinal Francesco Maria del Monte, one of Caravaggio’s most influential patrons, and labeled the Badminton Estate version a copy “after Caravaggio.”Wildenstein Collection version of “The Lute Player” (image via Wikimedia Commons)Sotheby’s similarly characterized the Badminton Estate painting as a copy when it offered the work at auction in 1969 and 2001. In a 2001 catalogue listing, the auction house theorized that it had likely been painted by 17th-century artist Carlo Magnone, an assistant to painter Andrea Sacchi. The work sold for £71,000 (about $129,883 today).The buyer of the work, British art historian Clovis Whitfield, insisted that the painting “corresponded exactly” with a description in Giovanni Baglione’s 1642 biography of Caravaggio. “Baglione mentions minutely observed details such as the reflection on dew drops on the flower,” Whitfield told the Guardian.In the text, Baglione wrote that Caravaggio “painted a youth playing a lute, and everything seemed lively and real, such as the carafe of flower filled with water, in which we see clearly the reflection of a window and other objects in the room, while on the petals of the flowers there are dewdrops imitated most exquisitely.”But according to the Guardian, Christiansen of The Met wrote to Whitfield’s late collecting partner Alfred Bader in 2007 that “no one — certainly no modern scholar — has ever or ever would entertain the idea that your painting could be painted by Caravaggio.”In a statement to Hyperallergic, Sotheby’s stood by its attribution of the painting as originating from the circle of Caravaggio and not the artist himself.“While we follow keenly scientific advances in AI and broader technology, and collaborate with scientists who are pushing the boundaries in the technical study and understanding of paintings, we see no reason to question the way in which we researched and catalogued the painting we sold in 2001, nor do we have any reason to doubt the attribution to Caravaggio of the painting that was exhibited at the Metropolitan Museum of Art in 2013,” the spokesperson said via email.Art Recognition and the Metropolitan Museum of Art have not yet responded to Hyperallergic’s inquiry.