When two stolen İznik tiles surfaced in a Barcelona auction, the decisive case did not come from an algorithm alone. Türkiye’s culture ministry says its TraceART system helped locate and archive the online listings; experts then connected the objects to records of tiles stolen from the Adana Grand Mosque, while police and laboratory analysis examined alterations to their surfaces. The tiles returned to Türkiye in February 2026.
That sequence is an unusually concrete example of artificial intelligence entering the cultural-property field as a search instrument rather than a substitute for scholarship. UNESCO now includes TraceART in its toolbox of practices for combating online trafficking. The system monitors online sales with keyword and image detection, sends flagged listings to heritage specialists, and can analyze social-media videos that may promote illegal excavation.
Scale is the problem AI is being asked to solve
Online marketplaces have multiplied the number of objects, images and descriptions that investigators must review. A specialist may recognize an unusual motif or suspicious provenance, but no team can manually inspect every listing in real time. Automated matching can narrow that field.
TraceART is designed around that division of labor. UNESCO describes it as part of a wider enforcement network linking heritage experts with customs authorities, prosecutors and police. A machine-generated flag is therefore the beginning of an inquiry, not its conclusion.
In the İznik case, the official account is careful about what followed detection. Archival documents, stolen-object records, scientific and technical examination, and a Catalan police laboratory report were all part of the evidentiary chain. Later analysis found that overpainting had altered the tiles’ appearance and confirmed the original decoration beneath it.
What this means for artists and collections
For artists, collectors and institutions, the development suggests that online visibility now has a second life. Images posted for sale can be compared against cultural-property records at scale. That may make it harder for stolen objects to disappear behind new descriptions or cosmetic changes, but it also raises questions about false matches, documentation quality and who can challenge an automated flag.
The strongest model is transparent human review. Systems should preserve the listing that triggered concern, record how a match was produced, and keep the final judgment with qualified experts and legal authorities. A visual resemblance is not provenance; it is a lead.
Artists can draw a practical lesson from the same process. High-quality photographs, measurements, dates, ownership documents and distinctive-detail images make a work easier to identify if it is stolen or misrepresented. Good records improve both conventional investigation and machine-assisted searching.
TraceART does not resolve the ethical and legal complexity of restitution. It does show how AI can help experts see a market that has grown too large to watch unaided. Its value depends less on the novelty of detection than on the rigor of everything that happens after a possible match.
ATIM approaches artificial intelligence with curiosity, independence, and balance. We explore its extraordinary potential to empower artists while examining the questions it raises around authorship, copyright, consent, creative identity, transparency, and human expression.
We are neither advocates for AI nor opponents of it. Our focus is how the technology is used—and what that means for artists.
Above all, our commitment is to artists. We look at every development through the lens of what can support, protect, and empower the creative community—identifying opportunities, examining potential risks, and providing artists with the information they need to make informed decisions about their work, their rights, and their future.