Artificial intelligence proves that she is ready too

Artificial intelligence proves that she is ready too

“Vostok, I’m Sneg 02. On the highway we have to turn left, damn it!” No need to go any further. Go on the defensive. It’s over. “” Yug 95, did you contact your superior directly? Warn him that the highway is being bombed. Don’t go in the column! Be careful! ”A radio broadcast between several Russian soldiers stationed on the Ukrainian front, intercepted last month on an unprotected canal, depicts panicked and confused comrades retreating chaotically after being caught under artillery fire. Away from the forehead, an artificial intelligence listened to her.

Open source front

The replica exchanges were automatically recorded, transcribed, translated, and analyzed by an algorithm developed by the American company Primer, according to Primer already sells artificial intelligence algorithms trained to transcribe and translate phone calls, but also to recognize predefined terms or key phrases.

CEO Sean Gourley says that with the invasion of Ukraine, the company’s engineers modified these tools to perform four new tasks: to identify and collect military broadcasts from web streams that broadcast their interceptions (lots of unprotected troop transmissions have been and continue to be published, translated and analyzed on various online channels (as well as other sources of information, including smartphone videos or social media posts); eliminate all background noise; transcribe and translate the discussions into Russian; highlight key statements relevant to the situation on the front. In some cases, this conversion involved the retraining of machine learning models to recognize colloquial terms for vehicles or weapons, for example.

It is this ability to train and retrain artificial intelligence algorithms along the way that will become a critical advantage in future wars, Gourley believes. Asked if the software is currently used outside the company, he said yes, but declined to provide details. While several other US companies have acknowledged that they have made technology, information and know-how available to Ukraine, CEO Primer is adamant: “We will not say who uses it or for what.”

The fact that some Russian troops use radio channels also surprised military analysts. “It seems to indicate an operation with limited resources and insufficiently prepared,” says Peter W. Singer, a political scientist and analyst with the think tank New America, which specializes in modern warfare. “Russia has used open communications interceptions to monitor its enemies in past conflicts, such as Chechnya, so all of them should have known the risks,” he added.

“It’s an indication of communications equipment failures, some arrogance, and possibly despair at the higher levels of the Russian military,” said Australian General Mick Ryan, a specialist in advanced military strategy and technology.

Calder Walton, assistant director of the Harvard Kennedy School’s Applied History Project, says the Russian invasion of Ukraine shows how valuable open-source information has become for agencies. “We are at an absolute point of reference in terms of the nature of information gathering and its availability,” says Walton, adding that the war in Ukraine has highlighted the importance of “digging” for relevant data in the most diverse and unconventional sources.

One of the examples given by analysts is the large number of Russian senior officers killed in the conflict, officers identified by Ukrainians using images captured not only by specialized satellites and drones, but also by digital cameras or smartphones.

Myth vs Technology

Estimates from the end of March speak of the elimination of at least 15 senior Russian officers, including 8 generals – more than were killed in the nine years of conflict in Chechnya, as well as in the campaigns in Afghanistan, Georgia and Syria.

Commenting on the balance sheet, Jeffrey Edmonds, a former director for Russia in the National Security Council (NSC) – the main body that assists the US president in matters of national security and foreign policy – told The Washington Post that Ukrainian forces appear to be targeting “Any person with gray hair standing near antennas is a sign that it could be a troop commander.”

Oleksii Arestovich, an adviser to Ukrainian President Volodymyr Zelensky, recently explained that the Ukrainian military has focused its efforts on “slowing down” the Russian invasion, in part by “decapitating” advanced command posts, adding that “excellent intelligence services and numerous Russian vulnerabilities” were the key to the success of this type of operation. Russian soldiers, for example, began using their mobile phones to communicate with each other, revealing not only their dissatisfaction and low morale, but also details about future or ongoing missions.

Author of a forthcoming book on the last 100 years of Russia’s (counter) intelligence war with the West, Calder Walton says that the NSA, the main US intelligence agency, and GCHQ, its British equivalent, have the most probably versions of the algorithms created by Primer.

Another algorithm used in Ukraine was developed by the American facial recognition company Clearview AI. “I remember watching videos of captured Russian soldiers that Russia claimed were actually actors. I thought then that if Ukrainians could use Clearview, they could get more information to verify and confirm their identities, “said CEO Hoan Ton-That in a BBC interview.

The company’s app can instantly identify someone just from a photo, and is already being used by the United States police and federal agencies to solve criminal cases. For this, explains Hoan Ton-That, the company’s artificial intelligence algorithm uses a database of 20 billion images collected from the public web, including from “Russian social sites such as VKontakte”.

The company offered its services to the Ukrainian government free of charge, and the application has so far been used by five government agencies in Ukraine (and) to identify dead soldiers in order to inform their families about their fate. The move is part of a campaign by which, according to a post on Telegram by Deputy Prime Minister Mikhail Fedorov, Ukraine wants to “dispel the myth of a special operation in which there are no recruits and no one dies,” bringing the cost to the attention of the Russian public. so tragic of the conflict.

The unseen faces of war

Facial recognition has advanced in strength and accuracy in recent years, becoming more and more accessible to the public. While Clearview AI says it makes its database exclusively available to law enforcement, other facial recognition services that identify using web images, such as PimEyes or FindClone, are available to anyone who wants to pay for them. .

Giorgi Gobronidze, an investor in Tbilisi, Georgia, who bought PimEyes in December 2021, recently said he had banned Russia from using the site after the invasion began, citing concerns that it would be used to identify Ukrainian targets. “We don’t want our service to be used for war crimes,” he said.

However, the involvement of private technology companies in armed conflicts also raises questions, with critics warning that they could take advantage of a crisis to expand relatively unhindered and that any mistakes made by software or its users could have serious consequences. in a war zone.

Evan Greer, deputy director of the advocacy group Fight for the Future, goes even further and says that facial recognition technology should be banned around the world because governments have used it to persecute minority groups and suppress dissent. Russia and China, among others, have implemented advanced facial recognition in city surveillance cameras. “Companies like Clearview are eager to exploit the humanitarian crisis in Ukraine to normalize the use of their malicious and invasive software. War zones are often used as testing grounds not only for weapons but also for tools used later to monitor the civilian population or to control crowds in the name of law enforcement, ”said Evan Greer.

Albert Fox Cahn, founder of the Surveillance Technology Oversight Project, also believes that the introduction of facial recognition in war could be disastrous, even if Ukraine uses it to tell the truth to Russian citizens. “This is a growing human rights catastrophe. When facial recognition makes mistakes in peacetime, people are arrested. When he makes mistakes in a war zone, innocent people are shot, “he told the US edition of Forbes magazine.

In response, Australian Hoan Ton-That says war zones can be dangerous when there is no way to distinguish enemy fighters from civilians. “Facial recognition technology can help reduce uncertainty and increase safety in such situations,” she said. US government-funded tests show that the Clearview algorithm has an accuracy rate of 99.85%, she added, which she says will “prevent misidentification in the field.”

Clearview is currently facing several lawsuits in the United States for infringement of rights, and the company’s use of photographs of people without their consent has been declared illegal in Canada, the United Kingdom, France, Australia and Italy.

However, the general view is that, in the long run, the collection and analysis of data using algorithms could become essential for battlefield operations. A U.S. military program, for example, called the Tactical Intelligence Targeting Access Node, proposes the creation of a ground station capable of interpreting information from sensors and other data sources, including smart weapons, in conflict zones.

If the Russian invasion of Ukraine was largely based on old tactics such as tank maneuvers and artillery bombardment, future wars could be based on new technologies, including artificial intelligence algorithms.

But at the same time, the use of artificial intelligence in military conflicts could turn into a game of mouse and cat, experts say, with efforts to confuse or mislead algorithms becoming just as important. The CEO of Primer also says: “Our philosophy of artificial intelligence and defense is that whatever algorithm you use to go to war, it will not necessarily be the one you end up with.”

This article appeared in issue 139 of . magazine.

PHOTO: Getty