In a historic shift for the art market, Germann Auction House in Zürich has sold a painting authenticated solely by artificial intelligence (AI), a first for the industry. The watercolor, an untitled piece by Russian artist Marianne von Werefkin, fetched nearly $17,000—almost double its estimated high of $9,000—on November 23.
This groundbreaking sale marks the beginning of what could be a new era in art authentication, as AI systems are increasingly used to verify the authenticity of works.
The auction house partnered with Art Recognition, a Swiss-based AI authentication company, to verify the von Werefkin piece. Unlike traditional methods, which rely on expert connoisseurs and physical analysis, Art Recognition’s “standalone AI” uses curated datasets to analyze patterns, brushstrokes, and other artistic traits. The AI generates a probability of authenticity without human judgment in its final evaluation.
Art Recognition’s CEO, Carina Popovici, emphasized that AI can enhance authentication processes, especially in challenging cases involving artists with disputed or fragmented catalogs, such as Amadeo Modigliani. She stated that while AI should not replace human experts, its use in conjunction with traditional methods could transform the art market.
At the same auction, two other works—by Louise Bourgeois and Mimmi Paladino—were authenticated by both AI and human experts, selling for $31,500 and $21,500, respectively. However, the von Werefkin piece stood out as the only work authenticated exclusively by AI.
The winning bidder, who chose to remain anonymous, praised the AI’s ability to identify patterns that confirmed the artwork’s authenticity despite its atypical style. He noted, “To be skeptical is good, but to be open-minded... In the end, the question is whether this method is accepted by the ‘relevant community.’”
Fabio Sidler, an auctioneer at Germann, said the sale showcased the effectiveness of Art Recognition’s algorithm and expressed confidence that this technology would encourage broader adoption in the art market.
While proponents hail this move as a milestone, skeptics warn against over-reliance on AI. British art historian Bendor Grosvenor expressed concerns about the limitations of AI when used independently, arguing that nuanced historical and contextual factors often require human expertise.
Popovici acknowledged these concerns, adding that collaboration between AI and human experts is crucial. However, she believes the sale could mark a shift in art market practices, with AI gaining acceptance for its ability to detect patterns in authenticity that might elude the human eye.
The integration of AI in art authentication signals a potential paradigm shift, offering faster, data-driven insights into provenance. Yet, as the art world balances innovation with tradition, the debate over AI's role continues. Germann Auction House’s pioneering sale invites both celebration and scrutiny, setting the stage for the evolution of authentication in a rapidly digitizing market.