Germany Funds €600,000 NEXUD AI App for Antiquities

Germany’s federal authorities have announced a €600,000 grant to NEXUD, a follow-up project to the ILLICID study. The NEXUD project is intended to find uses for and expand the capabilities of an image recognition app designed for use by police and customs and a webcrawler program. NEXUD follows on the ILLICID project, funded by the German government in 2015 with €1.2 million. The new project was announced by Prof. Dr. Markus Hilgert, an archaeologist and general secretary of the Cultural Foundation of the German Federal States in Berlin, which will lead the project. Dr. Hilgert headed the earlier ILLICID project as well.

Hilgert says the NEXUD project will enable authorities to identify whether an object may have been illegally looted or taken from an archaeological site using photographs analyzed by an artificial intelligence computer program. The criteria for such a determination will be submitted by “several dozen experts,” according to Dr. Hilgert.

However, what image recognition programs do is identify similar objects by their appearance, not tell their age or when they were found.

Partial view of the Mandelbrot set. Created by Wolfgang Beyer with the program Ultra Fractal 3. GNU Free Documentation License, Version 1.2 or any later version.

With sufficiently accurate and detailed data, image recognition and AI should be able to link image “types” to a range of potential countries of origin and to a database of those countries laws. However, connecting countries of possible origin to laws cannot tell anyone whether an object was looted yesterday or legally sold before the laws were enacted, or if it came into circulation 100, 300 or 500 years ago. Every country today has different heritage laws, created at different dates and enforced at different levels.

Determining an object’s legal status requires knowing its history of ownership and trade. The global extent of trade and the fact that ancient cultures did not exist within today’s geopolitical boundaries makes the determination of an object’s legal status even more challenging.

The NEXUD visual recognition program appears more likely to reinforce the erroneous idea that a machine can do a better job than an art historian and to promote the idea that most antiquities in circulation are unlawful. Since its developer says it will not be able to identify fakes, it is useless for that as well.

Is the NEXUD project actually going to benefit anyone? Wouldn’t educational programs for law enforcement have far greater capacity to increase their expertise – and provide them with more nuanced and valid answers to questions about illicit trade?

As British Museum expert St John Simpson, who assists UK authorities to identify illicit antiquities, told CPN in 2020:

“We’ve been shown images of objects by individuals and organizations where there was good reason to think that these objects were in Syria or had been recently exported from Syria but in every single case when I sat down with somebody and looked at these images they were all fake. The level of disappointment for law enforcement is quite palpable. You’ve got these very serious people hoping to press charges and we say actually all we can tell you is that it’s not genuine. It’s not worth any money, and there is no income stream coming from it.”

St John Simpson Interview: Afghanistan repatriation, Daesh, remote-archaeology and the ILLICID report, CPN, April 26, 2020.

Media Fails to Examine the Project’s Likely Results

One iteration of the inverse Julia transformation, 13 July 2009, author Prokofiev, GNU Free Documentation License, Version 1.2 or any later version.

Art trade organizations and museums condemn looting. Policy makers that support museums say the same. But there are compelling arguments that data is misused and false stories promoted on illicit trade that are harmful not only to the legitimate art trade but to a public interest in the legal circulation of art.

A recent article in The Art Newspaper, How do you spot a looted antique? Germany brings in team of experts to help, August 10, 2021, announced the NEXUD project without raising the most pressing questions about its purpose, its ability to deliver valid results, or the underlying policy agenda it would serve.

A response to the article by Vincent Geerling, Chairman of the International Association of Dealers in Ancient Art (IADAA) was also published in the print Art Newspaper, but is unavailable online. With Mr. Geerling’s permission, CPN reproduces it in full:

“The German government awarding a €600,000 grant to follow up the €1.2m ILLICID project relating to illicit antiquities raises more questions than it answers. Both projects are connected to the very restrictive German Cultural Property law passed in 2016, aimed at fighting against Germany’s status as an “El Dorado of the illegal cultural artifacts trade”, as claimed by Germany’s Federal Criminal Police Office (BKA) in 2014.

ILLICID launched on the assumption that “profits from illicit trade in cultural goods are an important pillar of organized crime”, while in the same sentence acknowledged that no reliable data existed to show this. In other words, it had come to a conclusion even before investigating. Neither ILLICID nor the German government has ever shown any evidence to support these claims.

At the time, I called the law “a solution looking for a problem”. Well, five years later, the problem still has not been found, and soon the law will be evaluated.

From Culture Minister Monika Grütters’ comments that the NEXUD project will provide “reliable answers from the academic community”, it seems clear that market experts will not be involved, yet they handle far more objects than academics so are mostly in a better position to assess them. The trade has offered to help with this, but so far has received no official response.

The Image-recognition software (already funded by the German government to the tune of €500,000) cannot detect fakes or illicit items, so alerts will depend on the team of experts. Will they be vetting everything offered for sale, and how will they limit the ‘web crawlers’ to the German market? With all this, how can their reaction be “swift” as described? If not, what’s the point of this exercise?

Dr. Hilgert, who will lead the project after heading ILLICID, is quoted as saying that trade is six times what he previously thought. Is that illicit trade, legitimate trade or web-based trade, as the article does not say? The article quotes him as saying that the ILLICID report’s main findings were that available information is “inadequate to ascertain whether items offered [on the internet] for sale are being traded legally or illegally”. If so, on what did he base his new figure, and what is it in absolute rather than comparative terms?

The only reliable data on the scale of the problem of illicit trade in cultural property in Germany comes from the German government itself. In official answers to parliamentary questions on March 2, 2021, it stated that from the introduction of the Cultural Property Act on August 6, 2016 until the end of June 2020, Germany made a total of 61 cultural property seizures on suspicion of import or export violations – 15 a year on average across the whole of Germany’s 16 states.

As far as we know, not a single case of terror financing has been found. However, as cases are investigated at state rather than federal level, the government says it has no idea of the outcomes.

By any measure, this is not a huge problem, so why throw so much money and resources at it?

Three simple questions:

• How many items among the 386,500 ILLICID studied were conclusively looted and trafficked? As far as I know, this has never been answered.
• How many of the 61 seizures investigated across four years proved to be valid?
• How many cases of terrorist financing (finding this out is one of the main goals of the German law) have been uncovered in the past five years?

Surely these and other questions regarding methodology, technology and expertise should be answered before committing significant sums to yet another project of this type?

Meanwhile, despite it knowing since last October that the figure was false, UNESCO continues to promote the value of illicit trade in antiquities as $10 billion a year as part of its equally bogus advertising campaign, The Real Price of Art. The fictional $10 billion number is oft-quoted to justify clamping down on legitimate trade, appearing in numerous media and academic articles even as I write.”[i]

Vincent Geerling, Chairman of the International Association of Dealers in Ancient Art (IADAA)

How is ILLICID being misused? In drafting a new UNESCO International Code of Ethics for Dealers and in the EU Action Plan on Trafficking in Cultural Goods.

Example of fractal art. The picture is reminiscent of a semipermeable membrane. June 2020, 19:56:27, author PantheraLeo, 1359531, CC0 1.0 Universal Public Domain Dedication.

Regrettably, the data from the ILLICID project appears to be used most today to misrepresent the extent of the illicit trade in antiquities. This is not just misleading, but harmful, because exaggerations about the extent of the trade misdirect resources that could better be used to preserve heritage at risk. And, as noted above, linking inaccurate data about the art trade with false claims about terrorism has resulted in unnecessary and burdensome EU legislation.

Few observers appear to be reading ILLICID accurately. A number of speakers at a September 14 UNESCO conference, ‘The Fight against the illicit trafficking of cultural property: for a strengthened global dialogue’ repeated erroneous interpretations of data from the ILLICID project. Margaritis Schinas, Vice President of the European Commission, said it showed that 98% of the ancient objects in trade were “illicit.”

These misstatements appear to be a factor in drafting forthcoming EU policies, notably revisions to the EU’s 1999 International Code of Ethics for Dealers in Cultural Property and the mysterious but allegedly comprehensive EU Action Plan on Trafficking in Cultural Goods – in which the antiquities trade organizations have been completely shut out from consultations and development. Despite requesting to participate, the art trade has been excluded from EU planning that will directly affect their industry, just as they were not invited to participate in the September 14 UNESCO Conference that was advertised as bringing all parties into the discussion.

What the ILLICID study actually concluded was that in all but 2% of cases, there was not enough provenance information available on the items being offered for sale to meet the documentary requirements for importation under the strict requirements of the new German law. Markus Hilgert stated as much in the Art Newspaper:

“The main findings of our ILLICID report were that the information available on the internet is inadequate to ascertain whether items offered for sale are being traded legally or illegally.”

However, documentation of provenance is not typically published in advertisements on the Internet. The UNESCO speakers reversed Hilgert’s statement to interpret the absence of documentation as proof of illegality.

Again, an illuminating quote from St John Simpson in 2020 below shows how the ILLICID data raises a completely different question – do its results complement other evidence showing that looted antiquities are not coming to the West when compared to the first Iraq war:

“The key fact about the ILLICID report is that out of the sample of the 356,500 objects that they say they looked at, only 0.1% were identified by them as genuine pieces from Syria or Iraq. That is very telling. I think it implies that there is a considerable amount of hyperbole about the volume of trafficking from those countries at least, even though there has sadly been a lot of destruction across the country. That hyperbole is something that we’ve seen in a lot of press stories over the last few years. Compare that to the fact that in the last nine years, we’ve not seen a single antiquity entering Britain or passing through Britain from Daesh-controlled areas of either Syria or Iraq.”

“That’s despite the fact that the UK is a major transport hub and there has been very high-risk profiling of individuals and freight from Daesh-controlled areas. We have also yet to see any objects on the market which we can show might have been stolen from the Mosul Museum in 2015.”

“I think that the absence of any such objects that we can demonstrably prove comes from these past nine years of conflict is very telling. It’s in complete contrast to the decade-long period from 1993-94 to 2003-2004 in Iraq where the satellite imagery of the ground matched the market picture. At that time there was organized looting of sites known to produce cuneiform tablets, glass, seals, coins, bullae and those sorts of things and we had huge quantities of those things on the market. That’s also a period when there was looting at museums and museum pieces were identified on the market. So in Iraq, in that decade, there was a complete match between the different data sets, whereas in Syria and northern Iraq in the last 8 years there’s a complete mismatch. That is something that needs to be better analyzed and explained. And I think that’s where future scrutiny ought to lie.”

What did the €1.2 Million ILLICID find out about illicit trafficking and terror?

Mandelbrot 5 two pairs of snails. Image rendered from my program called Manjira (マンジラ) 2.4. By Trinhhoa, public domain.

How did ILLICID fulfill its original goals and what are the expectations of this new project? According to the initial UNESCO announcement of the ILLICID project:

“The plan is to develop new analytical instruments and research approaches to better identify trafficking networks and processes and to better understand financial flows in the field of organized crime and terrorism, as well as to develop a digital object depository to be used by law enforcement and customs authorities.”

The project analyzed 356,500 items appearing in online art, coin, and other auctions of cultural materials between July 2015 and June 2017. The initial target area of war-torn Syria and Iraq was expanded to Ancient Cultural Objects from the Eastern Mediterranean (AKOM).  What did the researchers find?

• There were 6,133 AKOM items online out of the more than 350,000 total, approximately 1.7% of the total from Syria, Iraq, Greece, Turkey, Egypt, Cyprus or other eastern Mediterranean countries.
• Of these the 6,133 “relevant objects” found, the data indicated that only 24% of the “potentially” Syrian or Iraqi items were unquestionably authentic, 61.5% had insufficient information to tell, 12% were suspected fakes, and 2% were incorrectly classified.
• 71%, of items and lots started at under €1000 (10% at under €100). Higher value items were largely Greek and Roman. Only 52.9% of items sold at all.
• Of the 6,133 Eastern Mediterranean items, ILLICID identified 2,387 as “potentially” Syrian or Iraqi; of these, only 853 were offered as single items.
• There were huge number of fake antiquities being sold. Of the 6,133 “potentially” Syrian or Iraqi objects, probably only 24% of were unquestionably authentic, 61.5% had insufficient information to tell, 12% were suspected fakes, and 2% were incorrectly classified.
• ILLICID found no links at all to financing of terrorism or organized crime.

What ILLICID did find, in the end, was that there are significant numbers of objects representing the detritus of civilizations and the leftovers of the bazaar that are still in circulation – objects that museums and galleries do not need or want but that appeal to a market that values them for their age and history.

The Internet is an excellent way for legitimate art and antiquities sellers with brick and mortar galleries to reach new customers, and as reputable dealers use it, the Internet can promote transparency and accountability by making the objects offered visible to the world. However, ILLICID also demonstrated that the supposedly extensive illicit market in recent Middle Eastern antiquities on the Internet is really a tool for the fraudulent sale of fakes to the unsuspecting and ill-informed. That is useful information from ILLICID, because although most people in the art world knew it already, it’s nice to have it proven.

The Committee for Cultural Policy thanks St John Simpson once again for his illuminating interview and Vincent Geerling for granting permission to republish his letter to the Art Newspaper.

[i] Letter to The Art Newspaper from Vincent Geerling, Chairman, International Association of Dealers in Ancient Art (IADAA)