Masterpiece or cheap copy? Art historians and AI may not agree
Masterpiece or cheap copy? Art historians and AI may not agree
https://www.cnn.com/2026/03/09/style/ai-art-recognition-authentication
Publish Date: 2026-03-09 05:01:00
Source Domain: www.cnn.com
- Swiss AI firm Art Recognition disagreed with art historians about the authenticity of three Caravaggio paintings, asserting an 86% probability that the version at Britain’s Badminton House is genuine, contrary to the consensus that it was a copy.
- Art Recognition uses advanced AI models trained on comprehensive datasets of an artist’s undisputed works and known forgeries to determine authenticity.
- Skepticism remains among art experts regarding AI’s ability to replace traditional verification methods, which require deep historical and contextual knowledge.
- The AI’s process involves machine learning and computer vision techniques to train recognizing unique stylistic markers. However, full transparency about how these conclusions are reached is still a point of contention.
- The credibility of AI verification is a question in the high-stakes art market, where buyers rely on trust more than absolute certainties in authenticity.
- Despite resistance from traditional institutions, Art Recognition aims for its models to complement, not replace, conventional methods, potentially aiding provenance research and exposing forgery.
