AI and Old Masters: The New Frontier in Art Authentication

Machine learning is revolutionizing the art world, helping to authenticate paintings and attribute them to renowned artists. But can AI truly replace the human eye?

In the realm of art, artificial intelligence (AI) is making a significant impact, even in the more traditional areas like the Old Masters trade. This groundbreaking technology will be a hot topic at the upcoming Tefaf art and antiques fair in Maastricht, where Carina Popovici, CEO of the Swiss-based AI company Art Recognition, will unveil how they recently attributed a painting to a Renaissance German artist using AI.

Founded five years ago, Art Recognition has developed an AI system that offers precise and objective evaluations of an artwork’s authenticity. The company boasts over 500 completed authenticity evaluations, including the verification of contested works like an 1889 self-portrait by Vincent van Gogh at the National Museum in Oslo.

Art attribution is crucial in the art world, as confirming the authorship of a work can significantly increase its value and contribute to the field’s scholarship. For instance, “The Adoration of the Kings,” initially offered at auction in 2021 as a Circle of Rembrandt work, was later attributed to the Dutch master himself and sold for £10.9 million at Sotheby’s in December.

AI excels at pattern recognition, making it adept at identifying distinguishing features in an artist’s work. However, human reasoning remains essential, as AI struggles to grasp context. Jo Lawson-Tancred, author of the forthcoming publication “AI and the Art Market,” explains that while AI can flag paintings that do not fit an artist’s pattern, it cannot replace human judgment.

Carlo Milano of Callisto Fine Arts, London, emphasizes the importance of the data fed into AI systems. He warns that using questionable data can lead to dubious conclusions. Milano believes that AI will provide more information and reduce the margin of error, but it will never completely replace hands-on experience.

Conservators and art professionals remain skeptical about whether AI can account for factors like varnish, wear, or damage and if it will ever supplement or replace the human eye in judging a work of art.

Last year, Art Recognition found itself embroiled in a controversy over a painting known as the de Brécy Tondo, believed to be by Raphael. While an analysis by two UK universities using AI-assisted facial-recognition software concluded that the work was by the master, Art Recognition disagreed, stating that the de Brécy Tondo is not by Raphael with an 85% probability rating.

The Raphael dispute has since broadened, with different AI programs producing different results. In December 2023, a team led by scientists from the University of Bradford presented findings in a peer-reviewed paper, suggesting that the face of Joseph in Raphael’s “Madonna della rosa” may not be by the artist. However, Popovici challenges these findings, citing concerns about the training data used in the AI program.

Ugail, director of Bradford’s Centre for Visual Computing, maintains that his algorithm recognizes authentic works by Raphael with 98% accuracy. He emphasizes that AI is not replacing human expertise but rather providing another tool in the art authentication process.

Art historian Bendor Grosvenor acknowledges AI’s potential but highlights the need for better inputs in current AI attribution programs. He concludes that AI in this field still has room for improvement.

As AI continues to evolve, its role in art authentication will undoubtedly become more prominent. However, the debate over whether AI can truly replace human expertise remains open, leaving art enthusiasts and professionals alike to ponder the future of art authentication.

Source: Financial Times

---

Grow your business with AI. Be an AI expert at your company in 5 mins per week with this Free AI Newsletter