Art Handling in the Age of AI: A Personal Perspective from an Art Historian

There’s a newly unveiled mid-17th-century portrait of Dorothy Hobart, initially attributed to the “Circle of Peter Lely” in a local auction. After further examination, it seems this piece could be the work of John Michael Wright, a somewhat obscure but talented artist from London born in 1617. This portrait highlights the ongoing struggle for art historians in identifying and authenticating works, especially as artificial intelligence continues to grow in capability.

The challenge of what the future holds can be concerning. There’s a constant dialogue today asking what roles will survive when AI advances further. Personally, I’ve decided that I’d like to be an art handler. The thought of entrusting priceless works to a drone doesn’t quite sit right with me. After all, it takes a human touch to really appreciate and protect these cultural treasures.

Living in Scotland and frequently transporting art to London has refined my skills in moving paintings, putting me in an interesting position. This past year, I used two large artworks as teaching tools at the Royal Academy (RA) for my connoisseurship course, fitting them into the boot of my car. With ample cardboard and bubble wrap, I’m confident I could manage an entire gallery's worth of art on the road. Clarkson’s humor aside, it’s no small feat to navigate the complexities of transporting valuable pieces.

This indulgent humor, however, sheds light on a serious issue. The logistics involved in moving art have become increasingly complicated and cost-prohibitive, resulting in a significant decrease in loan exhibitions, which are vital for museums. In the 1990s, it seems we had a workable balance that has since eroded.

Returning to the portrait of Dorothy Hobart, I believed this painting could serve as an intriguing test for my students. Presented during our first session, their immediate inclination was to pull out their phones—eager to leverage technology for identification. However, Google Lens returned only the same generic attribution: “Circle of Lely.” Unfortunately, AI’s capacity for art identification proved limited in this instance, suggesting a range of possibilities without pinpointing Wright. In a stroke of irony, one AI response encouraged a student to “ask an expert”—namely, me. AI may guide us, but its occasional fabrications remind us of its limitations.

Consequently, we relied on traditional methods of inquiry. Our first stop was the National Portrait Gallery to examine examples of 17th-century portraiture, followed by the Heinz Archive, which houses thousands of photographs of British portraits. In conjunction with our list of names, the students and I compared features among Wright’s contemporaries. The moment of discovery was electric: a student identified a Wright portrait sharing a remarkably similar pose and background, leading us closer to an answer.

To wrap up our course, I organized a session with Dutch art historian Carla van de Puttelaar, who is currently working on a monograph dedicated to Wright. She confirmed my suspicions that our portrait most likely originated from this artist.

As a final note, both van de Puttelaar and I advocate for a dedicated exhibition showcasing Wright's work. Remarkably, he hasn’t had a solo show since 1982 at the Scottish National Portrait Gallery. If anyone’s considering a return to display, I know someone well-suited to handle it.

The experience of identifying, analyzing, and appreciating art remains an essential human endeavor, even in an age where technology offers promising assistance. While I humorously ponder my future role, it’s clear that the nuanced understanding achieved through human engagement, as distinct from AI's suggestions, cannot be replaced. This importance will always remain present as we navigate an ever-complex field of art and history.