How AI Is Helping Ukraine Anticipate Russia's Military Strategy

AI applications are helping Ukraine analyze intelligence on Russian tactics and strategy. 

Tools such as natural language processing are speeding the application of knowledge in real time, giving the users a significant advantage in allocating scarce resources against a better armed foe. JL 

Will Knight reports in Wired, image Gleb Garanich, Reuters:

The use of natural language processing technology to analyze Russian military communications is novel. As soldiers spoke, an AI was listening. Their words were automatically captured, transcribed, translated, and analyzed using several artificial intelligence algorithms. The ability to train and retrain AI models on the fly will become a critical advantage in war. AI advances could provide critical information more quickly, allowing military decision makers to outfox their foes. But battlefield use of AI may also become a game of cat-and-mouse, with efforts to deceive or mislead algorithms becoming just as important.

A RADIO TRANSMISSION between several Russian soldiers in Ukraine in early March, captured from an unencrypted channel, reveals panicked and confused comrades retreating after coming under artillery fire.

“Vostok, I am Sneg 02. On the highway we have to turn left, fuck,” one of the soldiers says in Russian using code names meaning “East” and “Snow 02.”

“Got it. No need to move further. Switch to defense. Over,” another responds.Later, a third soldier tries to make contact with another codenamed “South 95”: “Yug 95, do you have contact with a senior? Warn him on the highway artillery fire. On the highway artillery fire. Don’t go by column. Move carefully.”

The third Russian soldier continues, becoming increasingly agitated: “Get on the radio. Tell me your situation and the artillery location, approximately what weapon they are firing.” Later, the third soldier speaks again: “Name your square. Yug 95, answer my questions. Name the name of your square!”

As the soldiers spoke, an AI was listening. Their words were automatically captured, transcribed, translated, and analyzed using several artificial intelligence algorithms developed by Primer, a US company that provides AI services for intelligence analysts. While it isn’t clear whether Ukrainian troops also intercepted the communication, the use of AI systems to surveil Russia’s army at scale shows the growing importance of sophisticated open source intelligence in military conflicts.

A number of unsecured Russian transmissions have been posted online, translated, and analyzed on social media. Other sources of data, including smartphone video clips and social media posts, have similarly been scrutinized. But it’s the use of natural language processing technology to analyze Russian military communications that is especially novel. For the Ukrainian army, making sense of intercepted communications still typically involves human analysts working away in a room somewhere, translating messages and interpreting commands.

The tool developed by Primer also shows how valuable machine learning could become for parsing intelligence information. The past decade has seen significant advances in AI’s capabilities around image recognition, speech transcription, translation, and language processing thanks to large neural network algorithms that learn from vast tranches of training data. Off-the-shelf code and APIs that use AI can now transcribe speech, identify faces, and perform other tasks, often with high accuracy. In the face of Russia’s numerical and artillery advantages, intercepting communications may well be making a difference for Ukrainian troops on the ground.

Primer already sells AI algorithms trained to transcribe and translate phone calls, as well as ones that can pull out key terms or phrases. Sean Gourley, Primer’s CEO, says the company’s engineers modified these tools to carry out four new tasks: To gather audio captured from web feeds that broadcast communications captured using software that emulates radio receiver hardware; to remove noise, including background chatter and music; to transcribe and translate Russian speech; and to highlight key statements relevant to the battlefield situation. In some cases this involved retraining machine learning models to recognize colloquial terms for military vehicles or weapons.

The ability to train and retrain AI models on the fly will become a critical advantage in future wars, says Gourley. He says the company made the tool available to outside parties but refuses to say who. “We won’t say who’s using it or for what they’re using it for,” Gourley says. Several other American companies have made technologies, information, and expertise available to Ukraine as it fights against Russian invaders.

The fact that some Russian troops are using unsecured radio channels has surprised military analysts. It seems to point to an under-resourced and under-prepared operation, says Peter W. Singer, a senior fellow at the think tank New America who specializes in modern warfare. “Russia used intercepts of open communications to target its foes in past conflicts like Chechnya, so they, of all forces, should have known the risks,” Singer says. He adds that these signals could undoubtedly have helped the Ukrainians, although analysis was most likely done manually. “It is indicative of comms equipment failures, some arrogance, and possibly, the level of desperation at the higher levels of the Russian military,” adds Mick Ryan, a retired Australian general and author.

FEATURED VIDEO

A Timeline of Russian Cyberattacks on Ukraine

Most Popular

• 

  SECURITY

  Russia Inches Toward Its Splinternet Dream

  CHRIS STOKEL-WALKER

• 

  GEAR

  With Peloton Guide, the Fitness Company Bets Big on Body Tracking

  LAUREN GOODE

• 

  GEAR

  The Apple Watch Series 7 Is At Its Lowest Price Ever

  BRENDA STOLYAR

• 

  CULTURE

  WIRED’s Picks for the 15 Books You Need to Read This Summer

  WIRED STAFF

• 
  

ADVERTISEMENT

Calder Walton, a historian of espionage at Harvard, says the invasion of Ukraine shows how valuable open source information has become for intelligence operatives. Facial recognition software has been used to identify some individuals in videos of the conflict. “We are at an absolute watershed in terms of the nature of intelligence collection and what’s available,” Walton says. The conflict has highlighted the importance of mining different sources of intelligence. For instance, Ukrainian troops may have successfully targeted a number of Russian generals by looking for gray-haired individuals near antennas in satellite, drone, or other imagery. Russian troops have also taken to using cellphones, sometimes revealing their location and details of missions, as well as their frustrations and low morale.

Walton, the author of a forthcoming book on the past 100 years of Russia’s intelligence war with the West, says the NSA, the primary US signals intelligence agency, as well as GCHQ, the British equivalent, most likely have versions of the kinds of tools that Primer is using. But Primer is one of a growing number of companies that could make these technologies more accessible to those in the defense world and in private industry. The involvement of private companies in the war in Ukraine, such as those that provide satellite communications and imaging, raises questions about the power this gives those companies, and how they may become embroiled in an international conflict.

Tapping into open source intelligence data means sifting immense quantities of information. “The amount of open source intelligence is impossible for anyone to process,” says Emily Harding, a senior fellow at the Center for Strategic and International Studies, a policy research nonprofit, who wrote a report in January 2022 about using AI to mine open source intelligence. Harding says the intelligence world has gotten better at analyzing imagery using machine learning tools. Classifying the contents of images was one of the first tasks that demonstrated the power of modern AI. Harding says that Primer has distinguished itself for its ability to parse language.

Recent progress in AI could see the technology become a more powerful tool for analysis of text and speech. In recent years AI has become capable of summarizing or answering questions using text thanks to a particular kind of large machine learning model known as a transformer. This type of AI model is better able to make sense of input such as a long string of words in a sentence. Transformers have yielded AI programs that are capable of generating coherent news articles or even writing computer code to perform a given task.

However, Harding notes that the intelligence community will have to wrestle with the same problems that bedevil deployments of AI elsewhere, such as algorithmic bias caused by poor quality or unrepresentative training data. “It’s basically crap in, crap out,” she says. And because machine learning algorithms often work in opaque ways, Harding says, intelligence operatives will need to find ways to build trust in the conclusions these programs draw. An incorrectly transcribed communication could, of course, have deadly consequences on a battlefield, such as sending soldiers into harm’s way or misdirecting a missile strike.

Gathering and analyzing data using AI could eventually become central to battlefield operations. The US military is investing millions to develop AI software capable of ingesting and analyzing different signals in the field. A US Army program called Tactical Intelligence Targeting Access Node proposes creating a ground station capable of ingesting and drawing insights from many different battlefield sensors and data sources. If, for the most part, Russia’s invasion of Ukraine has relied on centuries-old tactics such as tank maneuvers and artillery bombardments, future wars that the US and other countries are preparing for may rely heavily on new technologies including AI.

Such advances could provide critical information more quickly, allowing military decision makers to outfox their foes. But battlefield use of AI may also become a game of cat-and-mouse, experts say, with efforts to deceive or mislead algorithms becoming just as important. “Our philosophy on AI and defense is that whatever algorithm you go into the war with is not going to be the one that you end up with,” Gourley says.