You will also be interested
[EN VIDÉO] Algorithms and their decision making Increasingly, algorithms are replacing human beings in decision-making or task performance. They take into account a number of parameters but the question of the transparency of the algorithms arises: can they explain their decision?
There is the work of the international community on open source intelligence (OSINT), as Futura referred to with facial recognition to identify Russian soldiers (see previous article below). There are also the powerful tools of facial recognition to provide security on the ground in Ukraine. Among the professional tools there is that of the American Clearview AI, a system favored by the American authorities and which can be tested in a demo version. The search engine is also one of the tools operated by the French company Tactical Systems Académie which has made it possible to identify Russian and Chechen fighters in Ukraine. It advantageously complements the results of the PimEyes tool used by the OSINT community.
Facial recognition is a way of giving names to the dead on the Russian side and dispelling doubts about possible false attributions. And precisely, to help the Ukrainian army, Clearview has just given it free access to its powerful search engine. During an interview with the agency Reuters, the boss of the company insisted that Clearview will not be marketed in Russia. On the other hand, the engine has already indexed more than 2 billion images from the Russian social network VKontaktethat is, the equivalent of Facebook in Russia. That database should be used by Ukrainians to identify the dead, even if the face of the deceased has been damaged.
Clearview is a powerful facial recognition tool. It was used to identify the perpetrators of the Capitol assault in Washington, by supporters of Donald Trump. The tool is often attacked by states in the European Union because of its ability to “suck up” photos from social networks, without any consent. © Clearview
Clearview AI regularly condemned in Europe
It seems that only the decomposition of bodies alters facial recognition according to research carried out for the United States Department ofenergy. The tool could also help identify russian agents trying to infiltrate via the numerous checkpoints guarded by the volunteer territorial defense battalions. But by the Clearview boss’s own admission to the news agency, you should not rely solely on this solution as the sole source of identification.
The Clearview boss also told Reuters the technology could be used to aid tracing and reunite separated refugees with their families. For the moment, the leader confessed that he was not sure how the Ukrainian Ministry of Defense was going to use the technology.
Still, if the technology seems effective, it requires going to dig into the image databases of social networks. Although this process is similar to how search engine indexing works google was refused by Meta and therefore Facebook and Instagram. For its part, Italy has just condemned the company Clearview AI to a fine of 20 million euros for having illegally collected and processed the personal data of Italians without a legal basis. Similarly, in December, the CNIL French has put it on notice to stop exploiting the aspiration of photos and videos. It remains to be seen whether VKontakte will let it go when this social network has a database of around 10 billion photos and is run by relatives of the Kremlin.
French company identifies Russian soldiers in Ukraine using facial recognition
In one hour and thanks to freely accessible facial recognition tools, the French company Tactical Systems Académie was able to identify a Chechen fighter sent to reinforcement Russian troops on Ukrainian battlefields. One example among many others of what an Internet user can do from home to document this war.
Article by Sylvain Biget, published on 03/11/2022
With a cellular connection remaining operational in virtually all of Ukraine and thousands of videos and photos posted in real time on social networks, the war in Ukraine is fertile ground for the development of what is known as theOSINT (Open Source Intelligence), i.e. open source intelligence. With the data available and a few tools available to everyone, anyone can reveal information about the loss of equipment or men, or the location of units and areas bombed by the Russian army. This is what made the site famous. Bellingcatthe work of the investigators of which has made it possible to establish that it is indeed a missile Russian who shot down the flight MH17 of Malaysian Airlines over the Donbass in 2014, and to identify the perpetrators involved in this tragedy.
One of OSINT’s tools during this conflict relies on facial recognition. Some services are accessible to the general public, such as FindClone, which makes it possible to recognize the faces of certain Russian soldiers and find them on the Russian equivalent of Facebook, VKontakte. This is how French society Tactical Systems Academy was able to identify in about an hour, Hussein Mezhidov, a Chechen commander sent to the Ukrainian front. The facial recognition of FindClone made it possible to find a photo of this fighter near Chechen President Kadyrov on his Instagram account. This French company, which works in particular for the French army, has also used several other tools, such as the face search engine PimEyes for confirmation. Asked by Wiredone of the leaders of Tactical Systems Academy specified that this research carried out by the OSINT community makes it possible to follow the movements individuals and that if they know they are being hunted, this could prevent them from committing war crimes.
On its Twitter feed, Tactical System explains how the team managed to precisely identify a Chechen commander involved in the war in Ukraine alongside the Russian army. © Twitter
Watch out for look-alikes!
While this facial recognition work was previously reserved only for the intelligence services, today everyone can do it from home. The problem is that authentication is sometimes risky and some people can be misidentified by OSINT enthusiasts. In this case, and since the data is shared openly, this can have consequences for people and their families who have nothing to do with the target. This is why, to counter the errors of the algorithms of facial recognitionthe use of several research tools as well as the attention of an experienced human eye are essential.
But these facial recognition tools also make it possible to limit certain acts of propaganda. Thus, a Ukrainian media recently claimed that a fighter pilot captured by the Ukrainian military was identified in a photo while alongside Vladimir Putin and Bashar al-Assad in Syria in 2017. Using a powerful reconnaissance engine developed by MicrosoftBellingcat concluded that it was not him, even though there was a strong resemblance.
Interested in what you just read?