Why A.I. is a useful servant but a dangerous master when it comes to cultural heritage
Artificial Intelligence is undoubtedly useful, especially when carrying out research, but it is also a minefield of fake news if you don’t do your homework properly.
As an example, take a report inspired by the recent dramatic theft of Royal jewels from The Louvre and published by Global Initiative Against Transnational Organized Crime. Titled Lessons from The Louvre, it includes the following statement: “Organized crime groups are increasingly targeting art and antiquities held in European collections, drawn by the continent’s cultural repositories and art markets.”
Whether this claim is true or not, the article provides no supporting evidence. It does mention a series of crimes that have taken place within Europe in recent years that may have been – even probably were – carried out by criminal gangs. Two of those mentioned even involved antiquities, although the rest did not. What they do not prove in any way is the veracity of the statement about such crimes being on the increase in Europe.
Now comes the A.I. part.
As an experiment, the Antiquities Forum asked A.I. the following question: “Is it true that organized crime groups are increasingly targeting art and antiquities held in European collections, drawn by the continent’s cultural repositories and art markets?”
The response was unequivocal: “Yes, organized crime groups are increasingly targeting art and antiquities held in European collections. The continent is a key hub for the illicit trade due to its wealth of cultural artifacts and active art markets, which organized criminals exploit for profit and other illicit purpose.”
It then provided a series of paragraphs under the heading Key reasons for the increase.
This all looked very convincing until further investigation showed that the conclusions were based entirely on sources including the article mentioned above, where significant claims had been made but without giving the evidence to show they were true.
Checking out the sources
In fact, in the case of the headline claim about organized criminals increasingly targeting European collections, the top three sources quoted were:
- the Global Initiative article;
- a Financial Action Task Force webinar from July 5, 2023, based on the deeply flawed February 2023 FATF report Money Laundering and Terrorist Financing in the Art and Antiquities Market, whose evidence does not support its headline claims; and
- the About section of the United Nations Office on Drugs and Crime’s website, which gives no supporting evidence and does not mention Europe at all. Previous claims and reports relating to the topic by the UNODC have been shown to be wrong.
Further sources include a 2022 European Union Action Plan against Trafficking in Culture Goods. Its claims that trafficking is a ‘lucrative’ business and that “increasing global demand from collectors, investors and museums” is driving looting and trafficking are based on the existence of UN Security Council resolutions, the size of the legitimate art market and estimates by Europol.
The problem here is that UN Security Council resolutions are preventive measures based on perceived risk rather than on evidence of actual crime; the size of the legitimate art market (which has been shrinking in recent years, not growing) sheds no light at all on crime levels, let alone showing that they are rising; and, at its own admission, Europol has no data to support the claim made on its behalf, despite all the headline figures on arrests and seizures (but not convictions or confirmation of the goods being illicit) relating to Operation Pandora and the rest.
In other words, as is almost always the case with such claims from the EU, they are not based on facts, but on supposition.
How the claims just don’t stack up
It’s a similar tale with Interpol, whose Cultural Heritage Crime page also features as an A.I. source in this context. There Interpol’s significant (but unsupported) claim is: “Trafficking in cultural property is a low-risk, high-profit business for criminals with links to organized crime. From stolen artwork to historical artefacts, this crime can affect all countries, either as origin, transit or destinations.”
In fact, not one of the sources given by A.I. to support the robust (but misleading) conclusions it comes to stand up to scrutiny.
Unfortunately, as reports from many sources have shown, researchers looking for confirmation of their suspicions when it comes to the international art market often fall victim to confirmation bias, failing to check the sources they cite through footnotes and embedded links because on the surface they seem to support what they believe.
Because of this, it has been surprisingly easy to debunk long accepted, but false, claims about the art market. And it also explains why the Antiquities Dealers’ Association (ADA) and the International Association of Dealers in Ancient Art (IADAA), who between them support the work of the Antiquities Forum, for years have operated a policy of checking primary sources for claims wherever possible.They provide those sources wherever possible in promoting their own views and arguments, and ask their audiences to check the sources given for their own satisfaction. Unfortunately, this is a highly sensitive and controversial arena where no one – no matter who they are – can simply be taken at their word. And A.I. is not going to fix that any time soon.
Recent Comments