AI Goes to War
The lessons of Ukraine, the rise of the drones, the challenge of China
The Military AI Challenge
Artificial intelligence is fundamental informational technology that completely reshapes all other technologies, including military technology.
AI and predecessor and related technologies - machine learning, autonomous vehicles, UAVs, etc. - have been used for intelligence analytics, autonomy, and decision support in military applications.
AI technology continues further digitization of the military: Sensors collect all information possible on the battlefield. AI has been used to digest and analyze that information, turn it into intelligence, identify targets, and target weapons.
Incorporating AI and autonomous technologies into military platforms is a step-change in capabilities, based on the greater, speed, precision, and the cost-at-scale available. Billion-dollar platforms and centers of analysts can be replaced by million-dollar drone fleets with embedded AI algorithms.
The countries that win the AI technology race will have a technology advantage that will translate into economic and military supremacy.
The main competitor to the United States in AI is China. The US leads on account of US-based big technology firms dominating in AI R&D, while China is a strong number two in their influence and contribution in AI research as well in AI model developments. The EU and the UK also playing a role in advancing AI.
China’s advances in AI poses geopolitical risks, while EU contributions in AI support our interest as allies, and counterbalances rogue nations, which are laggards.
Ukraine As An AI Warfare Testbed
Recent conflicts, in particular the war in Ukraine, have showcased the disruptive effectiveness of AI and autonomous technologies, and their impact is growing as AI rapidly improves. Ukraine has been a test laboratory for AI warfare.
In the Ukraine-Russia war, Ukraine’s Secret Weapon has been Artificial Intelligence. Most visibly impactful has been Ukraine’s use of UAVs and drones. They have destroyed aircraft at airbases, taken out tanks, helped blunt Russian air superiority, assisted with artillery targeting and served as loiter munitions:
Early in the war, Turkish-made Bayraktar TB2 UAVs degraded Russian air advantage and disrupted Russia’s invasion forces.
Attaching obsolete anti-tank grenades to consumer-grade quadcopter drones enabled Ukraine to take out Russian heavy tanks with weapons costing $100.
Low-cost drones attacked an airbase in Russia and destroyed a TU-22 strategic bomber.
Ukraine used autonomous surface craft to attack Russian warships, disabling a ship by hitting it with a sea drone carrying 450 kg of TNT.
Saker scout drones in service by Ukraine can identify and target 64 types of Russian military objects, and carry a 3kg payload at a range of 10 km. They are now targeting and completing attacks autonomously.
While Ukraine’s inventiveness helps it beat Russia in the drone conflict, Russia too has been adapting, for example, using Iranian-made Kamikaze drones to attack targets. Both Russian and Ukrainian forces are integrating traditional weapons with AI, satellite imaging and communications.
AI in Operations
Ukraine’s AI advantage has gone far beyond drones. This conflict has become “software-defined warfare,” with AI playing its most important role in operations. Roles and Implications of AI in the Russian-Ukrainian Conflict, states, “A key role of AI in Ukraine’s service is the integration of target and object recognition with satellite imagery, prompting Western commentators to note that Ukraine has an edge in geospatial intelligence.”
Ukraine does this through the “Delta system.” Communicating in decentralized way and via starlink, this is a “network-centric warfare” of system of command, control, communications and intelligence.
This C3I “network-centric warfare” operational perspective is familiar to US military planners and thinkers, and the digitization of the military has been ongoing for decades, long before AI arrived.
What’s different is the introduction of AI has the potential to accelerate precision intelligence processing, analytics and tactical decision-making. Vastly greater data can be collected for greater battlefield awareness; decisions can be accelerated, making for a tighter OODA loop; greater lethality (at lower cost) can be brought to bear.
Palantir CEO Alex Karp said earlier this year that Ukraine war shows urgency of military AI:
"Your ability to identify the right technology and implement it will determine what happens on the battlefield. One of the major things we need to do in the West, is realize this lesson is completely understood by China and Russia."
USA’s Strategic Imperative - Intelligentized Warfare
Other nations are taking notice. Notably, Israel has been using AI in their conflict with Hamas. During this conflict, the US shot down a drone attack from the Houthis in Yemen.
Inspired by Ukraine, Taiwan is building drone fleets to match China for their national defense.
China has increased their spending on AI-related military technologies. They lead the world in drone technology specifically, and they are a strong second to the USA in Artificial Intelligence in general. Most importantly, they are pouring a ten times greater ratio of their military spending into these new technologies than the U.S.
A report called Offset-X, from SCSP a panel of experts headed by former Google Chairman Eric Schmidt, examined the challenge of maintaining America’s technological-military supremacy in the face of China’s challenge. Offset-X discusses the Chinese perspective:
The PLA has closely studied the “American way of war” … precision munitions-battle network warfare, which the PLA calls “informatized warfare.” It has identified the U.S. military’s command, control, communications, and targeting networks as the U.S. operational center of gravity and developed a theory of victory centered around system destruction warfare. This system destruction warfare concept aims to disrupt the flow of internal information, the time sequencing of control-attack evaluation systems, and essential components of an adversary’s operational system through kinetic and non-kinetic means to deconstruct and defeat a military.
China and the PLA plan to leap-frog the USA’s digital information-based warfare with “Intelligentized Warfare” capability that surpasses and defeats it.
The only way for the U.S. to counter that is to beat China to that future and build it ourselves. This future “Intelligentized Warfare” will have four features:
AI-driven: Use AI and high-velocity analytics in the C3I stack to drive high-capacity data gathering, intelligence analysis, real-time decision-making and force and weapons control.
Software-based and adaptive: Systems will adapt to conditions and results in the theatre of conflict. Software can be patched faster than new hardware manufactured, so platforms that are extensible will win.
Autonomous systems: UAVs, drones, autonomous ships and vehicles, and unmanned weapons platforms, will proliferate across a range of complexity, capability and cost levels. These systems are massively less expensive to build, deploy, and maintain than their manned counterparts. As AI improves, their cost-advantage and capabilities improve as well.
Informational warfare: Since the military system will run on data, warfare will extend to disrupting the enemies information networks at all levels. Jamming communications, destroying recon drones, counter-measures, cyber-security, and deception will be a part of war. The arena of conflict also includes propaganda, Deep Fakes and disinformation on social media.
Future Warfighting Weapons - Autonomous Systems
This report on the U.S.-China AI contest over autonomous weapons and killer robots makes key points about the weapons that will win the AI-based arms race. Some example weapons systems we are seeing already:
Ghost Sharks from Anduril are cheap and fast AI-powered autonomous subs. Costing less than a tenth of 1% of the cost of a nuclear sub, it has no pressure hull and is much smaller, avoiding the need for any crew, and can dive deeper than a manned sub. Australia will get Ghost Sharks by mid-2025.
A Black Hornet 3 drone weighs only 33 grams and can fit in the palm of a hand; it can fly almost silently for 25 minutes, sending video and high-definition images back to its operator. Multiple militaries are using these drones.
The FH-97A is a jet fighter-like drone unveiled by China last November. It will operate with a high degree of autonomy to provide intelligence and added firepower alongside manned combat aircraft.
Drones and autonomous weapons systems have many advantage and benefits:
Many useful roles: Recon drones replace ground-based artillery observers for recon or artillery spotting. You can arm a drone as a loiter munition. More advanced and faster drones can be used for broader data gathering or larger range attacks.
Low Cost: Unmanned aerial drones are in some cases so cheap that the weapons used to counter them are more expensive. Another example - the Switchblade drones are cheaper than a flying hour for a larger platform. Cheap drones are lethal enough to expensive manned aircraft that it forces those out of the sky, as they can’t compete in a war of attrition.
Expendable: Because they are cheaper and unmanned, the lower cost systems can be expendable or attritable. You can lose a system without losing a life. Drones can degrade enemies by forcing them to expend limited missile stocks for counter-measures. Robot scouts can be one-time-use to probe defenses, or can be used as “Kamikaze” weapons against high-value targets.
Lethal: Simple low-cost drones have greatly improved the lethality of artillery, rockets and missiles in Ukraine. Drones make manned aircraft vulnerable: attack and transport helicopters have become so vulnerable that they have been almost forced from the skies, their roles now increasingly handed over to drones.
Faster tempo: Armed drones can eliminate the distance between sensors and shooters, increasing the tempo of battle decision-making. They have potentially 24/7 up-time. Drones can be deployed in swarms to more quickly overwhelm enemies.
Flexible: Jet-powered drone fighters could perform maneuvers the human body wouldn’t tolerate, such as high-G-force turns. Aerial drones can also do away with the pressurized cockpits, oxygen supplies and ejector seats required to support a human pilot.
While we have described mainly flying drones, it applies also to autonomous sea vehicles, both subs and ships, and land-based autonomous weapons systems, both moving and fixed. All such autonomous drones and weapons, teamed with humans, are more flexible, faster, cheaper and better, and thus are poised to supplement or displace manned systems in many warfighting roles.
The extreme cost-effectiveness of autonomous military platforms - unmanned vehicles, robotic missile launchers, UAVs, unmanned subs - is so disruptive a shift, it will obsolete most prior warfighting capabilities. The cost differential alone will make prior generation platforms as untenable as horses were in the era of tanks or battleships after the arrival of aircraft carriers.
Part of the step-change will take humans out of war-fighting; humans are costly to train and support, expensive to protect, tragic to lose. Unmanned platforms will proliferate on land, sea and air. While previous generations of UAVs were remote-piloted, next generations will be further autonomous, capable of higher-level decision-making.
This will radically alter how battles are fought; battlefield strategy will need to follow.
The cost-effectiveness and accessibility disparity of such technologies will disrupt existing technology and future military investments. National defense spending priorities will need to change. The U.S. is responding. It will acquire and field “multiple thousands” of autonomous, unmanned systems within the next two years.
The AI Battlespace
As important as autonomous weapons are, the most important use of AI in warfighting will be its ability scale up high-velocity high-precision intelligence gathering, analytics and tactical decision-making. As stated in the report on the U.S.-China AI contest over autonomous weapons:
“Perhaps even more revolutionary than autonomous weapons is the potential for AI systems to inform military commanders and help them decide how to fight – by absorbing and analyzing the vast quantities of data gathered from satellites, radars, sonar networks, signals intelligence and online traffic. Technologists say this information has grown so voluminous it is impossible for human analysts to digest.”
We already rely on machine learning and AI to sift through voluminous data, to find key intelligent or identify targets. As AI competence and capability grows, we will further automate the data-to-intelligence-to-decision pipeline. Further automation at sufficiently high quality will scale into a much faster higher-precision military decision loop. Super-human AI could lead to super-human OODA loops.
Near-term AI will be able to manage the entire cycle of recon, acquire target, make weapons selection and make firing decision, plus manage logistics and strategic decisions. It will become a question of military ethics and grand strategy to determine where to put the human in the decision-making loop. We will likely keep the human officer as the responsible executive and the AI as the presenter of data, options and implementer via autonomous weapons systems for many tactical military actions.
High-precision leads to Microtargeting, utilizing high-precision information and high-speed decision loops to launch attacks on key combatants or commanders. Consider how the U.S. has used drones to attack terrorist leaders, or how Ukraine targeted senior Russian military leaders in the Ukraine conflict.
AI-based warfighting is software-based warfighting. AI and the software that runs most of the autonomous systems will be updateable and upgradeable far more quickly than developing a new weapons system. It will need to be to survive. The pace of AI and software technology change is increasing.
The digital sensor and AI capabilities of even lost-cost drones are such that total battlefield awareness and intelligence is possible. To defend from this, combatants will wage electronic warfare to disrupt the command, control, communications, and intelligence (C3I) of the enemy. The main battle will be a battle for informational awareness advantage. There will be counter-measures, cyber-warfare, spoofing, and jamming of communications, up to and including space.
Winning the battle for information superiority will be the equivalent of having air superiority in prior conflicts. For the superior combatant, the battlefield will grow more transparent. AI accelerates and compresses the decision-making processes. It makes survival on the battlefield more challenging for those at an information disadvantage.
Conclusion
“AI is increasingly powering warfare. Based on the rate of progress in the AI field, I predict that in ten years, it will be the dominant force.” - Alexander Wang, ScaleAI CEO
If information and AI superiority in tomorrow’s war is the equivalent to air superiority in yesterday’s war, it can help one understand why AI will be the dominant factor and force in military power in the not-too-distant future.
As AI advances dramatically in the next 10 years, it enables the automation of much of every human endeavor, including warfare.
The advanced military of 2035 will include large fleets of autonomous weapons platforms and systems orchestrated by a combination of AI systems processing of vast amounts of data and human coordinators and operators, working together at the tactical, operational, and strategic levels.
Such a military will be dramatically faster, more nimble, and may also be dramatically more cost-effective than today’s military.
As we said at the outset, China’s rise in AI poses geopolitical risks. The only path forward for the U.S. to maintain military superiority in the AI era is to both win the AI race and to transform our military into an AI-centric battle-fighting system, to develop “Intelligentized Warfare.”
PS. B-21 Bomber - AI in the Clouds
Not all future weapons will be unmanned. Earlier in November, the newest Air Force stealth bomber, the B-21 Raider, took its first flight. However, this nuclear bomber weapon does boast an array of AI-related and cloud-based digital technologies. According to some, this US bomber is why China suddenly wants to talk about nukes and AI.