AI helps to detect illegal cultural assets
Project KIKu launched – Fraunhofer SIT develops app to support cultural and investigative authorities
A key challenge in combating the illegal trade in stolen cultural goods is that illegally traded objects are difficult to identify. The KIKu project – a German acronym for Artificial Intelligence for Cultural Property Protection – aims to facilitate the work of the responsible authorities, especially customs and police, and is funded by the Federal Government Commissioner for Culture and the Media. To this end, the Fraunhofer Institute for Secure Information Technology SIT and cosee GmbH are developing an app that can use artificial intelligence to provide automated information on whether, for example, an antique vase or statue could have come from a looted dig or was otherwise illegally acquired. The project was presented on November 4th, 2020 to numerous relevant actors in the field of cultural property protection from Germany and the EU, who met at the invitation of the Minister of State for Culture Monika Grütters for the 7th EU CULTNET meeting in the context of the German EU Council Presidency.
In 2018, Interpol registered more than 91,000 stolen cultural objects worldwide and almost 223,000 objects seized by prosecution services, e.g. antique coins, ceramics, historical weapons or fossils. Most of these objects came from illegal excavations and were intended to be taken out of the country avoiding customs. In order to locate cultural goods from looted excavations or stolen cultural objects, customs officers search for illegally traded objects at airports or online auctions, for example.
An AI application may help the investigating authorities in the future in determining whether or not an art object may be exported or traded. Experts in IT forensics and media security from Fraunhofer SIT and specialists from cosee, a Darmstadt-based company specializing in digital product development, will be developing a prototype app that can support cultural and investigative authorities in evaluating cultural assets.
Minister of State for Culture Monika Grütters: "We want to use state-of-the-art technical means in the fight against illegal trade with cultural assets. Innovative applications with artificial intelligence help the responsible cultural and investigative authorities to distinguish quickly and reliably between legally and illegally traded cultural property in practice. With the development of the KIKu app, the Fraunhofer Institute for Secure Information Technology is making an innovative, sustainable and practically relevant contribution to this, which we are pleased to support with funds from the German government's national AI strategy."
Investigators can use their smartphone to take photos of the object to be checked from multiple perspectives. The KIKu app sends the images to a server, where a deep learning network checks the photos taken. With the help of artificial intelligence, it tries to detect similarities between the photographed object and already known cultural treasures. If the object is found to be stolen, the investigator receives a direct warning with the associated information. With the help of this similarity detection, AI can also be used to assess cultural objects that are not yet known to have been stolen from looted excavations. When the AI finds similar objects, it selects corresponding data sets and sends them to the KIKu app on the investigator's smartphone. Within seconds, these objects provide the investigator with an initial assessment of the region and era from which they may have originated, as well as an initial indication of whether the origin of the cultural object in question requires further investigation.
The Federal Government Commissioner for Culture and the Media has earmarked funding of up to 500,000 euros for the project from the Federal Government's national AI strategy. KIKu is a follow-up project to the transdisciplinary research project "ILLICID - Illegal Trade in Cultural Property in Germany," which was initiated and funded by the Federal Ministry of Education and Research in 2015 as part of the "Research for Civil Security II" program.