EPISODE · Aug 6, 2026 · 50 MIN
Noah Charney on AI art authentication and 85.7% Caravaggio
from Roborant Review
Can an algorithm tell you whether a painting is a real Caravaggio — and should we let it? And what happens when a performance artist steals Hitler's favorite painting and hangs it in an immigrant family's kitchen?In this episode, host Hugh Leeman sits down with Noah Charney — art crime historian, founder of ARCA (the Association for Research into Crimes Against Art), TED speaker, and author of dozens of books including The Art of Forgery and The Museum of Lost Art — for a wide-ranging conversation about theft, forgery, lost masterpieces, and the artificial intelligence now quietly reshaping how the art world decides what's real.Charney tells the extraordinary story of the artist Ulay, who in 1976 stole Carl Spitzweg's The Poor Poet from a Berlin museum as a political "aktion" — with Marina Abramović photographing and a getaway van idling in the snow — and how Charney came to record ten hours of interviews with Ulay in the final year of his life. He explains why two-thirds of all old master works may be lost to history, and why the ones we can visit aren't necessarily the ones that mattered most.Then he unpacks the "fourth pillar" of art authentication: AI image analysis. Using the real case of a $71,000 "follower of Caravaggio" painting that an algorithm flagged as the genuine article, Charney walks through how a closed, purpose-built AI is trained on a catalogue raisonné, fed forgeries and synthetic fakes to learn what not to be fooled by, and used as a tiebreaker when equally decorated scholars refuse to agree. He makes the case against the all-powerful single connoisseur — from Bernard Berenson's sales commissions to the Wacker Van Gogh forgery trial — and teases a coming "demotion" of a famous painting that he predicts will make headlines around the world.
Embed this episode
Ready to play
Noah Charney on AI art authentication and 85.7% Caravaggio
No transcript for this episode yet
Similar Episodes
No similar episodes found.