Artificial Authenticity

Vincent Van Gogh’s “Undergrowth with Two Figures”

The first time I saw “Undergrowth with Two Figures” at the Cincinnati Art Museum, I was awe-struck. Seeing the way Van Gogh’s thick, waving brushstrokes come to life when light dances across the canvas is an experience that a picture can’t replicate. It also doesn’t capture the humbling feeling of knowing that a master’s hands touched the canvas in front of me.

But what if they didn’t?

Experts estimate that somewhere in the neighborhood of 20% of museum holdings may be forgeries or misattributed works. The provenance of “Undergrowth” is not in question, but statistically speaking, I’ve probably had that exact transcendent feeling in front of a canvas I falsely believed to be touched by a master’s hand.

Wolfgang Beltracchi painted in the styles of modern masters so convincingly that museums and collectors believed they were looking at lost works. Alongside his wife, Helene, he created and sold hundreds of paintings attributed to artists such as Max Ernst, Heinrich Campendonk and André Derain. He sourced period-appropriate materials, aged canvases and built backstories that moved millions of dollars through the market before his arrest in 2010.

Before him, Elmyr de Hory moved his way through galleries around the world with convincing works in the style of Pablo Picasso, Henri Matisse and Amedeo Modigliani. Trained in Munich and Paris, his story eventually became the subject of a book, and was immortalized in Orson Welles’s film “F for Fake.”

Non-Physical Forgery

Deception is not new to the art world, but generative AI creates new pathways through which it can enter the market. AI is now being used to draft polished provenance letters, generate gallery websites with curated exhibition histories, fabricate collector timelines and seed online coverage that suggests buzz and legitimacy. This information is then used to take out insurance policies on artworks that never physically existed. It’s a new type of art forgery that requires neither painter nor painting.

While this may not add to the number of forged or misattributed works we see hanging in museums, it could change the historical record and cause artificial shifts in market values as more and more of our cultural living exists in digital space.

The art world’s existing authentication infrastructure was developed to catch human forgers working with physical materials. It is not yet equipped to flag a synthetic paper trail or a fictitious exhibition history that is seeded across dozens of legitimate-looking websites. Researchers are now developing digital provenance registries, AI-detection tools trained on synthetic documents, and cross-referencing databases for exhibition claims in an effort to keep up with the way criminals exploit new digital tools.

The Line Between Art and Content

Outside of deception by ill-intended characters, AI brings up other questions around authenticity in the art world. If an artist creates an artwork using some or all AI-generated content, is it deceptive not to disclose that contribution? In 2023, a photographer revealed that his winning entry at the Sony World Photography Awards had been AI-generated. The response sparked an ongoing conversation about whether AI outputs can be considered art.

Next month, we will dig into that question as we continue to explore the ways technology will shift the arts sector in 2026.

Jeni Barton is Director, Digital Products at ArtsWave. The monthly Arts and Tech column explores the benefits and challenges of technology in the arts.