The art industry’s lack of transparency, especially regarding pricing and final transactions, has been cited by many as a major factor preventing the market from expanding its collector base and reaching the new buyers it needs to sustain accelerating global growth. Many apps and startups have tried to address this problem, but none have succeeded in democratizing a sector that has become a highly visible lifestyle phenomenon while still operating at a level of secrecy and exclusivity.
Magnus Resch, the art market economist and entrepreneur, has built an entire career around the art market’s blind spots, writing extensively about the systems that determine artistic success, the sustainability of exhibitions, and the behavior of art collectors. His latest project is an update and re-release of an app called Magnuswhich describes itself as “Shazam for art.” Originally launched in 2013 as a regularly updated guide to art world events, the app went offline during the coronavirus crisis when museums and galleries closed their doors. Resch recently introduced a new AI-powered version, currently in beta, that builds on the original premise with image recognition and a large database of artwork prices. Take a photo of an artwork and, within seconds, you will receive information about the artist, title, and most importantly, price.
“The art market doesn’t have an attention problem. It has a conversion problem,” Resch told the Observer. “People visit art galleries and galleries, but many of them don’t buy because they don’t feel informed enough. We answer three simple questions: What is the price? Is it fair? And is it a good investment?” He sees a younger generation interested in art but unsure of how and when to buy, not only because prices are opaque and transaction costs are high, but also because the purchasing process can seem intimidating and exclusive.
The research that led to the redevelopment of his eponymous app underscored how fragmented information is in the industry, but also shed light on how value is created. He said there was a lot of hard work on the ground. For more than 15 years, he and his team have visited art galleries around the world to collect prices by hand. “This is data that cannot be accessed anywhere else. We have combined that data with auction results for over 50 years,” he added, stressing that the app provides users with access to “the world’s most comprehensive art price database,” with millions of gallery prices combined with secondary market sales data.
“We cleaned up [the data]“Linking duplicates and creating a unique identifier for each artwork,” Resch explained. In most databases, if a business is sold twice, you see two entries; Magnus offers a complete history of this artwork, including where it was offered, when it appeared at auction, sale prices and whether it has been previously offered by a gallery. Context, in this sense, is what sets Magnus apart from other apps and data aggregators. “It depends on the context: who the artist represents, where the work has been shown, which museum owns it, how similar works have been performed and how momentum develops over time.
As for why we’re bringing Magnus back now, recent advances in AI have dramatically improved the speed and accuracy of image matching. Recognizing works of art is relatively simple when someone is looking at a famous painting or sculpture, but is more difficult in the art market, where the works may be newly made, unique, undocumented, or visually similar to other works by the same artist. Artists also influence each other, Risch noted, and may have similar visual languages. “Training the model on a thousand Basquiat works does not mean that it can recognize a thousand and one. It must also distinguish it from thousands of visually similar works by other artists.”
Another challenge was how to trace identity, as there is no universal identifier for works of art that can be compared to the ISBN for books. “We had to build it ourselves,” Resch said, but once he and his team had a database of millions of uniquely identified artworks, the accuracy of matching Magnus improved dramatically. “An image search from Google, ChatGPT or Claude can often identify the artist’s name, but they never give you the price. We do.”
Discussions about the impact of AI on the art world typically focus on image generation and issues of authorship, but Resch sees AI as key to achieving market transparency. He insists that it is a working system, not a creator. “As we showed on our website Study scienceAn artist’s success depends less on the artwork itself than on reputation, visibility, and professional networks. AI may generate beautiful images, but it cannot replace the human relationships that create cultural and economic value.
Except perhaps for the relationship between conductor and counselor. “Magnus learns what users like and recommends artworks based on their taste, budget, and location,” Resch continued. (If someone constantly scans affordable abstract paintings that contain a lot of green, the app can recommend similar green works in galleries nearby or elsewhere in the world.) “No human advisor can continuously monitor the entire art market. AI can do that. This is a much more practical use of AI, it gives people access to information and discovery tools that were previously only available to insiders.”
Ultimately, what the AI-powered version of Magnus offers is information; What people do with this information is up to them. Resch acknowledged that more informed buyers can be more skeptical. “I think this is a healthy thing,” he said. “Transparency does not weaken the art market, but rather expands it.” “More people will buy art if they understand what they are looking at and feel like they are participating in an open system rather than a closed system. The art market doesn’t need more visitors. It needs more buyers. Transparency is how we get there.”
More on Artificial Intelligence in Art
