Authenticity: Verifying That a Work of Art Is Genuine

Between 2007 and 2024, more than 6,000 fake paintings flooded the market attributed to Norval Morrisseau, an Ojibwe artist who died in 2007. The forgeries were produced by multiple rings operating in Ontario. Financial losses exceeded C$100 million. In 2024, Ontario police called it "the biggest art fraud in world history." In 2025, the Morrisseau Estate partnered with a Montreal start-up called Acrylic Robotics to train an AI tool called Norval AI that can assess the probability that a given painting is authentic. The tool uses machine learning to analyze brushstroke patterns, pigment composition, and stylistic features. In July 2026, a Toronto-based authentication firm called Art Experts Canada Inc. announced it had removed more than 50 additional Morrisseau forgeries from the market so far that year, each permanently marked as fake. Read the report at CP24.

Authenticity in art is the verification that a work is genuinely by the artist to whom it is attributed. An authentic work is one that was created by the named artist, or in the case of workshop productions, created under the artist's direction and with their involvement. Authenticity is not the same as quality. A painting can be authentic and mediocre. It is also not the same as originality, since an artist may produce multiple versions of the same composition. Authenticity is a claim about authorship, and verifying it is one of the most difficult and contested activities in the art world.

This entry covers what authenticity means, how it is established, who makes authentication decisions, what happens when those decisions are challenged, and how AI is changing the field in 2026.

How Authenticity Is Established

Authentication relies on three categories of evidence: provenance, technical analysis, and connoisseurship. Provenance is the documented history of a work's ownership from the artist's studio to the present. A complete provenance with no gaps is strong evidence of authenticity, though provenance documents can be forged. In July 2026, a man named Andrew Crowley was sentenced to two years in prison for attempting to sell three stone Cycladic figures and an Anatolian statuette through Sotheby's with forged provenance documents. Forensic analysis determined that the printer technology used to produce the documents did not exist in the 1970s, as claimed, and the typeface was not developed until 2001. Read the report at The Art Newspaper.

Technical analysis examines the physical materials and structure of the work. Pigment analysis can reveal anachronistic materials: if a painting attributed to Vermeer contains titanium white, a pigment invented in the 1920s, the painting is not by Vermeer. X-radiography can show underlying compositions or structural features consistent with the artist's known working methods. Infrared reflectography can reveal preparatory drawings that match or contradict the artist's style. A 2026 study published in "Heritage" journal used multispectral imaging, Raman spectroscopy, and mass spectrometry to analyze Michele di Matteo's 15th-century altarpiece "Coronation of the Virgin," demonstrating how non-invasive techniques can characterize materials and support authentication. Read the study at Heritage.

Connoisseurship is the visual judgment of an expert who has studied an artist's oeuvre for decades. The connoisseur evaluates style, brushwork, composition, and handling, comparing the work to known authentic examples. Connoisseurship is subjective and can be wrong, but it remains a primary tool, especially for Old Masters where technical analysis may be inconclusive.

Forgery and Its Consequences

The history of art forgery is long. Han van Meegeren sold fake Vermeers to the Nazis in the 1940s. Tom Keating claimed to have produced more than 2,000 forgeries in the styles of Old Masters before his exposure in 1970. John Myatt and John Drewe sold forged paintings attributed to Matisse, Giacometti, and Chagall in the 1980s and 1990s, with Drewe tampering with archive records at the Tate to create false provenance. Shaun Greenhalgh produced forgeries of sculptures, paintings, and documents over 17 years before his arrest in 2006.

In 2026, federal prosecutors charged a father-daughter duo, Karolina Bankowska and her father, with producing at least 200 forged paintings attributed to artists including Andy Warhol, Banksy, and Andrew Wyeth. Bankowska walked into a New York City gallery carrying a painting signed by Wyeth and sold it on consignment. The gallery owner, Robert Rogal, told the Associated Press: "The provenance was a little fuzzy. But she seemed credible. It wasn't an obvious counterfeit." Read the report at AP News.

The financial stakes are high because authenticity determines value. A genuine Warhol silkscreen can sell for millions. The same image, if deemed inauthentic, is worth the cost of its canvas. This is why authentication disputes are among the most litigious in the art world. The Andy Warhol Art Authentication Board, which dissolved in 2011, was sued multiple times for declaring works inauthentic. The board's decisions could destroy millions of dollars in value overnight, and collectors who received negative rulings sometimes sued.

AI and the Future of Authentication

Artificial intelligence has entered the authentication field, with mixed results. The Swiss company Art Recognition, founded in 2019, uses deep learning models to analyze paintings and assess authenticity. The company has made headlines for bold claims: in 2021, it claimed that Rubens's "Samson and Delilah" at the National Gallery in London was a fake, a claim that leading Rubens experts rejected. In 2025, Art Recognition claimed that two versions of "Saint Francis of Assisi Receiving the Stigmata" attributed to Jan van Eyck were misattributed, with 91 percent certainty for the Philadelphia version and 86 percent for the Turin version. The Van Eyck expert Maximiliaan Martens of Ghent University disputed the findings, arguing that only about 25 paintings are attributed to Van Eyck, an insufficient dataset for training an AI model. Martens called for peer-reviewed scientific publications explaining the company's methods. Read the report at Artnet News.

The Morrisseau case shows AI's potential when deployed with institutional support. The Norval AI tool was trained on authenticated works from the Morrisseau Estate and is being used alongside traditional provenance research and connoisseurship. The estate's executive director, Cory Dingle, told The Art Newspaper that the tool "has grown to do many other functions that will help with museums seeking provenance as well as law enforcement, such as catching the person who painted the fraudulent work." Read the report at The Art Newspaper.

A 2026 article published in "Humanities and Social Sciences Communications" by Nature examined how generative AI is complicating the copyright landscape for authentication. The article proposed an output-oriented test based on "substantial recognisable linkage," asking whether an AI-generated work can reasonably be perceived as drawing on protectable expression in a prior work. This framework may reshape how courts evaluate authenticity in the age of AI-generated images. Read the article at Nature.

Who Decides?

Authentication authority is fragmented. For living artists, the artist's own statement is usually definitive. For deceased artists, authentication falls to several bodies. Artist foundations and estates often maintain authentication committees: the Basquiat estate, the Keith Haring Foundation, and the Pollack-Krasner Foundation have all operated authentication services. Catalogue raisonne projects, which compile complete records of an artist's work, serve as de facto authentication authorities because inclusion in the catalogue is itself a statement of authenticity. Auction houses employ in-house specialists who assess authenticity before sale, and their due diligence processes can trigger criminal investigations, as the Sotheby's case in 2026 demonstrated.

Independent authentication firms like Art Experts Canada Inc. have emerged to fill gaps. The firm's removal of 50 Morrisseau forgeries in 2026 shows how private authentication services can function as market regulators when institutional authorities are absent or overwhelmed.

Related Terms and Concepts

Authenticity is connected to several art world concepts. Provenance is the ownership history that supports authenticity claims. Attribution is the related process of assigning a work to a specific artist, which is the basis for any authenticity claim. The conservator provides the technical analysis that can support or undermine authenticity. The Old Master market is particularly vulnerable to authenticity disputes because these artists worked with large workshops where authorship was collective. For more on the market context, read our post on understanding art auctions or our guide on art as investment.

See Authentication in Action

Authentication is rarely visible to the public, but museum exhibitions sometimes reveal the process. The Rijksmuseum's Operation Night Watch, which began in 2019, has shown visitors how technical analysis can uncover an artist's working methods and support authenticity claims. The Metropolitan Museum of Art periodically publishes technical bulletins that document the scientific analysis of works in its collection. The Ghent Altarpiece restoration project has made its technical findings publicly available, including AI-assisted crack analysis.

For more on how the art world verifies what is real, read our entries on provenance and attribution, or explore our post on why provenance matters to learn more about the paper trails that make or break authenticity claims.