Can Machine Vision Replace Art Expertise?

Say the words "artificial intelligence" or simply, "A.I." in an art setting, and people think of either cutting-edge, new media art, or of misinformation., hallucination, and plagiarism. But there's a case to be made that those words should prompt you to think about very old art and about very new technology's use in finding out what's real.

My colleague at Artnet, Jo Lawson-Tancred has a new book out called A.I. and the Art Market, that serves as an accessible guide to a range of ways that artificial intelligence and machine learning are impacting the art market.

There's a lot in the book about valuing art, about selling art, and about navigating the intellectual property challenges around A.I., but we thought we'd drill down into the question of art authentication, which has drawn plenty of headlines and controversy in recent years, all on its own. After all, huge amounts of money hinge on the question of whether a given piece of paint on canvas is actually considered to be by a particular old master painter.

The art market has an entire robust world of art historical expertise built up around art authentication, which is revered, but sometimes also viewed with suspicion as corruptible and subjective. Then, here come various forms of A.I. art authentication with its own jargon and new kinds of suspicion aimed at it. So who should you trust? Jo has spent a lot of her time talking to various players to help begin to answer that question, and today we dig into the thorny question.