Is It? The Art Mystery Podcast

In Orlando a museum exhibited what it was claimed were a lost cache of 25 Basquiat paintings. Then the FBI raided and seized them. In this episode, we used cutting edge AI technology to check those 25 paintings to see if any might actually be by Basquiat, or were all of them forgeries? Joining us is art market expert Magnus Resch and Art Recognition founder, Carina Popovici.

This episode features a portrait of Eleonora de Toledo, the wife of Duke Cosimo de Medici and an impressive woman, indeed. But is this portrait of her by the Medici court painter Bronzino? Or by his pupil and devoted follower Alessandro Allori? To explore who made this painting, we spoke with Bruce Edelstein, a professor and researcher based in Florence who is part of the global network of NYU. He is also a Bronzino specialist and has written a book on Eleonora. We are also joined by Alessandra Conti from Art Recognition who explains the cutting edge AI image analysis and how to interpret the results. So...it is a Bronzino? Or an Allori? Tune in to find out!

In this episode, we’re turning our attention to the work of Mahmoud Sa'id, one of the most important and influential figures in modern Egyptian art. Trained in Europe yet deeply rooted in Egyptian culture, Sa’id created a body of work that blends academic technique with a distinctly local vision. His paintings occupy a central place in the history of twentieth-century Egyptian modernism—and they also raise important questions about attribution, variation, and legacy. To discuss a painting that she selected for AI image analysis, I’m joined by Jehane Ragai, emeritus professor of chemistry at the American University in Cairo. Professor Ragai is a pioneer in the scientific analysis of artworks, with decades of experience applying chemistry and materials science to questions of attribution, authenticity, and conservation. Her work has been especially influential in the study of modern Egyptian painting, where scientific evidence can play a crucial role in clarifying complex art-historical and market questions. She’s the author of two excellent books that are sitting on my shelf: "The Scientist and the Forger" and "Technical Art History." I’m also joined by Batu Arda Düzgün, AI Developer at Art Recognition, where he works at the intersection of machine learning, visual analysis, and art authentication. Batu develops systems that analyze artistic patterns at a level of scale and precision impossible for the human eye alone, offering new tools for understanding authorship, consistency, and deviation within an artist’s oeuvre. Let's ask the question: Is it a Sa'id?

This episode is about the Guidelines on the Ethical Use of AI for Image Analysis in Art Authentication. AI is everywhere, but that means that the term is being used with sometimes reckless abandon. For any serious field—like art attribution—it’s important to lay out recommendations for how best, most transparently, and most effectively to use AI. These guidelines were developed jointly by Art Recognition and the Center for Art Law (with a little help from me), and they offer a framework for how artificial intelligence can be used responsibly, transparently, and collaboratively in the authentication of artworks. To discuss the guidelines, I’ve invited Carina Popovici, co-founder of Art Recognition and a past guest on this podcast, and Irina Tarsis, the founder of the Center for Art Law. The three of us were the main drivers of these guidelines, with significant feedback and input from a variety of others, many lawyers working with the Center for Art Law.

Today’s featured picture is called Portrait of a Man. It certainly looks like a Velazquez. But you know what question I’m going to ask. Is it? To answer this question, I've invited two guests. First, Batu Arda Düzgün, originally from Turkey, is a researcher at Art Recognition, a company that uses artificial intelligence to help authenticate artworks. With a background that spans both computer science and art history, Arda is at the forefront of applying cutting-edge machine learning to some of the oldest and most mysterious paintings in the world. Joining him is British art historian Matthew Wilson, an expert on visual symbolism and the author of several widely-read books, including Symbols in Art, The Hidden Language of Symbols, and Art Unpacked, all published by Thames & Hudson. Matthew’s work uncovers the layers of meaning behind what we see, offering insight into the codes, allegories, and visual metaphors that artists have used across centuries. Arda’s work likewise uncovers layers of meaning but in the form of complex image analysis and the use of AI. What do they have to say about the artwork featured in today’s episode? Listen to find out.

Most episodes of this podcast will feature a special guest art expert who chooses a painting that they’d like to have analyzed—a painting with uncertain or debated authorship. We are joined by, an AI specialist who reveals the test results and we discuss them together. But here and there we’ll include an interview that is relevant to the topic of art authentication, art historical mysteries, and technologies like AI, but which steps outside of the usual format. This is one such episode. In it, I’ll be interviewing one of the founders of Art Recognition, the firm that sponsors this podcast and runs the AI image analysis tests—Carina Popovici.

Is it a Van Dyck? That’s the question we seek to answer in this episode. The painting in question is one of several versions of a Portrait of Don Diego Messia Felipe de Guzman, some of which are by Anthony van Dyck. But is this one? To answer this question, my guests for this episode include Prof. Dr. Nils Büttner, a distinguished art historian specializing in the visual culture of Germany and the Netherlands from the 15th to the 17th centuries; and Alita de Feudis, an art historian affiliated with Art Recognition, a Swiss technology company specializing in AI-driven art authentication.