America is betting on artificial intelligence

Victory in the armed conflict of the future will ensure superiority in high technologyMilitary experts from the leading countries of the world are constantly trying to predict what the war of the future will be and what weapons will be crucial to ensure victory in it.

The United States, one of the world's leading military powers, is also striving to solve this problem, faced with challenges that are not typical of armed conflicts of the past. Moreover, the military-technical potential of states that Washington considers opponents or strategic rivals has significantly increased.

"The American Armed Forces have long relied on technological superiority to ensure their dominance in the conflict and ensure the national security of the United States. However, in recent years, technology has developed rapidly and spread mainly due to advances in the commercial sector. As former Defense Secretary Chuck Hagel noted in this regard, such a development threatens to undermine the traditional sources of the United States' advantage in the military sphere," the report "New Military Technologies: Background and Questions for Congress", prepared recently by the Congressional Research Service, states.

This document is a kind of "educational program", a brief overview of the main directions and problematic issues characteristic of the work carried out today in the leading military powers of the world in the field of developing new military technologies and creating advanced weapons: artificial intelligence; deadly autonomous combat systems; hypersonic weapons; directed energy weapons; bio- and quantum technologies. At the same time, the list of countries by the authors of the document is limited to America and its two strategic rivals – China and Russia.

Artificial intelligence in uniformArtificial intelligence (AI; artificial intelligence – AI) is considered as one of the promising technologies that can radically change the very nature of the wars of the future: failure in this area will damage the national security of the United States.

However, as indicated in the report, so far even the term "artificial intelligence" has not been officially formulated at the state level in the United States. In general, experts understand AI as "a computer system capable of cognition at the human level."

There are three groups of AI: limited (narrow AI), general (general AI, also – artificial general intelligence / AGI) and superintelligent or super-AI (artificial superintelligence).

Limited AI (also narrow-purpose AI or weak AI) is able to solve only one task (problem). It can function in real time, but, being able to work only in a given range, it processes information only from a limited, narrow data set and does not possess any signs of human consciousness (mind).

General AI (also general-purpose AI or strong AI) is able to solve a wider range of tasks of different nature, including those for which it was not prepared in advance. In fact, this is a man-made version of the human mind, capable of solving any mental tasks that a person can do. He will be able to plan his actions, reason and make judgments (decisions) on a wide range of issues, including in a rapidly changing environment and in conditions of complete uncertainty. He will also be capable of self-study.

Finally, super-AI experts mean such a system, which, as Nick Bustrem (also Bostrom), director of the Institute for the Future of Humanity and professor of the Faculty of Philosophy at the Oxford Martin School (a research and policy unit based on the Department of Social Sciences at Oxford University), notes in his work "Superintelligence: Paths, Dangers, Strategies", according to its capabilities "significantly surpasses the cognitive abilities of people in almost all areas of interest." By and large, super-AI is the same "Skynet" and "Matrix" from the famous movie blockbusters. Moreover, there are fears that the emergence of super-AI may make the existence of a person simply "superfluous".

However, as stated in the report, "general and superintelligent artificial intelligence does not yet exist, and perhaps will not exist at all." But limited AI is widely used today in various fields, including military and special services, allowing to facilitate the work of a person or even replace him in a number of areas: surveillance and intelligence; ensuring the security of objects (persons); processing and analysis of intelligence and other data; conducting cyber operations; logistics; improving the effectiveness of the military system command and control of semi-autonomous and autonomous weapons systems, etc.

Systems equipped with such AI perform actions faster, process a larger data flow and allow the introduction of new ways of using forces and means. For example, using a swarm of drones. At the same time, the authors of the report consider the ability of AI to ensure the creation of increasingly realistic digital fakes of photos, audio and video ("deep fake" technology) to be particularly important: "Opponents could use these AI capabilities in their information operations within the conflict in the gray zone. Deep fake technology can be used against the United States and its allies to create false news reports, influence public discourse, undermine public trust and attempt to blackmail government officials."

At the same time, systems based on limited AI may be subject to algorithmic distortions as a result of improper use of their training data or models. The report indicates that cases of racial bias in facial recognition programs have been repeatedly recorded due to the lack of diversity in the images on which AI systems were trained. There are also cases of gender discrimination and other failures in the operation of limited AI algorithms for object recognition and classification. In the military sphere, such a flaw can lead to mistaken identification and unintentional killing of non-combatants by autonomous combat systems with AI elements.

The Pentagon is constantly increasing spending on the topic of AI: if in the 2016 fiscal year (fy.) $ 600 million was allocated for this in the unclassified part of the budget, then in 2023 it was already 1.1 billion, and the number of projects implemented by the department exceeded 685. In addition, to coordinate work in the field of AI, a Joint AI Center (JAIC) was created in the US Defense Ministry in 2019, which in January 2022, JAIC became part of the new Main Department for Digital Technologies and AI (CDAO), subordinate to the Deputy Minister of Defense.

In 2018, the US Department of Defense put into effect an AI Strategy, then adopted five ethical rules for the use of AI (responsibility, equality, traceability, reliability and manageability), and on May 21, 2021, Deputy Secretary of Defense Kathleen Hicks issued a memorandum on the responsible use of AI technologies. Finally, in June 2022, the US Department of Defense approved a Strategy and a plan for implementing the principles of responsible use of artificial intelligence technologies.

Artificial Intelligence CommissionIn 2018, the Independent National Security Commission on AI (NSCAI) was formed with the task of developing recommendations to the President and Congress on "promoting the development of AI, machine learning and related technologies to comprehensively meet the needs of US national security and defense."

In March 2021, she presented a report in which she comprehensively considered issues related to the use of AI and proposed to focus efforts on five main areas: investment in R&D; refinement of AI for national security tasks; recruitment and training of AI specialists; protection and development of technological advantages of the United States; establishment of global AI cooperation.

"Americans have not yet realized how deeply the revolution in artificial intelligence (AI) will affect our economy, national security and well-being. Much remains to be learned about the power and limitations of AI technologies. Nevertheless, important decisions need to be taken now to accelerate work in the field of AI for the benefit of the United States and to protect against the malicious use of AI," the preamble of the 756-page report stated.

A number of recommendations of the NSCAI commission have already been formalized in the form of laws, and the Law on National Defense for fiscal year 2022 instructs the head of the Pentagon to submit annual reports to Congress on the implementation of these recommendations within the Pentagon. This was done, apparently, not least because, as indicated in the report of the US Accounting Chamber published in March 2022, "Artificial Intelligence: the Ministry of Defense should improve strategies, inventory process and interaction guidelines," the Pentagon's activities in this area are still very far from ideal.

It should also be mentioned that in October 2021, the first AI strategy was adopted by the NATO bloc (among other things, it is planned to create a network of AI centers), and one of the main areas of cooperation between the United States, Great Britain and Australia within the framework of the AUKUS trilateral agreement defined "AI and autonomous systems" (plus hypersonic weapons, quantum and other advanced military technologies).

According to the report, the United States sees China as its main opponent in the field of AI. In 2017, the latter adopted the "Next-Generation AI Development Plan", in which AI is considered as a "strategic technology" capable of influencing the course and outcome of international rivalry.

At the same time, it is particularly noted that in the PRC in the field of AI there are practically no borders between commercial companies, research institutions, the government and the military.

Russia is also considered in the report as a potential opponent in the field of AI. Special emphasis is placed on the well-known words of Russian President Vladimir Putin that "whoever becomes a leader in this area will be the ruler of the world." However, at the same time they prefer not to mention the words said after that: "And I really would not like this monopoly to be concentrated in someone's specific hands." It is clear why – after all, America is just striving to achieve global superiority in the field of AI. Including the military.

Americans are particularly concerned about the work being carried out in the Russian Federation on weapons systems that have AI elements and high autonomy of actions: automated combat modules for land and marine equipment, semi-autonomous and autonomous robotic combat systems of all types based, including those capable of operating as part of a swarm. The United States is also concerned about our efforts to introduce AI into communication and control systems, electronic warfare, work with social networks, etc.

At the same time, the document notes that a number of experts, based on the volume of open academic publications of Russian experts on AI issues, consider it unlikely that Moscow will achieve a significant breakthrough in this area. Probably, they somehow forgot about the fact that there may also be closed publications. The authors of the report understand this: "Other analysts object that such factors may be insignificant, arguing that although Russia has never been a leader in the field of Internet technologies, it has managed to become a noticeable force in cyberspace."

When robots killThe next advanced military technology that requires close attention is lethal autonomous weapon systems (lethal autonomous weapon systems – LAWS), by which the US MOD understands a class of weapon systems capable of detecting a target and attacking it with their weapons regardless of the operator's actions.

Moreover, the directive of the Ministry of Defense No. 3000.09 specifically indicates the difference between LAWS from those autonomous combat systems in which the operator has the ability to monitor the use of his on-board weapons and, if necessary, stop the attack, and from semi-autonomous combat systems that attack only the target selected by the operator.

The creation of LAWS became possible only thanks to the development of appropriate AI elements and means of detection and targeting, which are able to provide an offline solution to the task of detecting a target, classifying it, developing a solution for the use of on-board weapons systems, followed by the issuance of targeting data, then attacking the target and evaluating its results, and, if necessary, making a decision on repeated attack. Suart Russell, a computer science professor at the University of California, even described such weapons systems as "the third revolution in military affairs after gunpowder and nuclear weapons."

American experts consider it a priority to use such systems in conditions where there is no stable communication, and, thus, the possibility of using classical means of destruction (aviation, missiles, reconnaissance and strike drones, etc.). It is also believed that LAWS systems can improve accuracy and reduce the likelihood of "friendly fire" or the destruction of civilian objects and non-combatants.

However, the document states that experts from about 30 countries and 165 non-governmental organizations are sure of the opposite: in their opinion, such weapons systems pose an increased danger for technical (hacking of on-board control systems by the enemy, AI malfunctions, etc.) and ethical (in fact, the machine makes its own decision to kill people) reasons, Therefore, it is necessary to prohibit the development and application of LAWS systems at the international level.

The authors of the report claim that they do not know whether work is underway in the United States to create LAWS and whether they have been adopted by the Armed Forces, but note that there is no official ban on such actions in the country. Moreover, the Directive of the Ministry of Defense No. 3000.09 clearly states: LAWS should be created and delivered to the troops in the future, but with the obligatory condition of compliance with "the laws and customs of war, applicable treaties, rules of security of weapons systems and applicable rules of warfare." It is also required to ensure that commanders and operators are able to "exercise an appropriate level of human judgment about the use of force."

At the same time, the latter does not mean the mandatory inclusion of the operator in the control loop of LAWS systems, but ensuring more active human participation in deciding where, how and why LAWS should be used in combat. Plus, it is necessary to ensure that the weapon system can exit the battle or request the operator's help if it is unable to independently perform a combat task in a time corresponding to the intentions of the head of the operation or the commander in the Theater of operations.

According to the Pentagon, China is actively working on the creation of LAWS systems. For example, former Defense Minister Mark Esper noted that a number of Chinese weapons companies, such as Ziyan (developer of unmanned aircraft systems), position their developments as capable of autonomously searching, classifying and defeating targets.

Chinese experts themselves define LAWS as a weapon system that has at least five characteristics: the presence of weapons on board; the autonomy of actions during the entire process of completing the task; the impossibility of withdrawing the system from a combat mission; the indiscriminate nature of actions (the system can kill or injure regardless of conditions, scenarios and type of targets); the ability to evolve in environmental conditions (including in a way that exceeds human expectations).

Invulnerable HypersoundHypersonic weapons are another element of the new revolution in military affairs.

In the basic version, it includes samples capable of developing a speed of M = 5 or more. However, since ICBM warheads also have such speeds, it is customary to refer to the considered combat means those that can perform a controlled flight on the entire or significant part of the trajectory at hypersonic speed and, most importantly, can perform maneuvering, including anti-aircraft, during the flight to the target.

There are two types of combat vehicles: hypersonic cruise missiles (hypersonic cruise missile) and hypersonic gliding aircraft or gliders (hypersonic glide vehicle). The former are equipped with a propulsion system, with the help of which they follow the target after launching from an air, sea or land-based carrier, and the latter plan to reach the target after dropping from a rocket or carrier aircraft.

The main advantages of combat hypersound are high speed and accuracy, as well as maneuverability, which gives unpredictability of the flight path. This reduces the effectiveness of countering existing air defense/missile defense systems, but at the same time, as experts point out, poses a threat to maintaining strategic stability. Moreover, hypersonic weapons of destruction today do not fall under any of the regimes for the control of strategic weapons.

"The US military claims that modern means of detecting ground and space echelons (missile defense) are not capable of detecting and accompanying hypersonic weapons," the report "Hypersonic Weapons: Background and Questions for Congress", prepared by the Congressional Research Service on October 27 this year, states.

"Hypersonic weapons make it possible to strike time-critical and very important targets. They challenge enemy detection and interception systems and complement the capabilities of existing cruise and ballistic missiles," General Mark Milley, Chairman of the US Armed Forces Joint Staff, stressed at a hearing in the Senate Committee on Armed Forces Affairs on March 4, 2020.

However, a number of experts believe that the role of hypersonic weapons is overestimated, since Moscow and Beijing already have high-precision nuclear ICBMs at their disposal, the large-scale use of which the American missile defense system is not capable of reflecting. However, as well as their massive strike with the use of hypersonic weapons: the Russian Avangard, Dagger and Zircon, and the Chinese DF-41 ICBMs with the DF-ZF nuclear hypersonic glider, and in the future – the Starry Sky-2 hypersonic complex.

The leading roles in this area today are played by China, the Russian Federation and the USA. Moreover, the Pentagon, which not so long ago almost closed the work on combat hypersonic as unpromising from the point of view of creating a usable weapon in the foreseeable future, has again intensified this activity, since such weapons appeared in the PRC and the Russian Federation.

As a result, since fiscal year 2019, expenditures on combat hypersound in the United States have gone uphill: in 2020, $ 2.6 billion was included in the military budget for these purposes, in 2021 - 3.2 billion, in 2022 – 3.8 billion, and for fiscal year 2023, the Pentagon has already requested 4.7 billion. In addition, the Missile Defense Agency has been receiving funds for the creation of a hypersonic weapon protection system since 2017. So, for 2023, it is planned to allocate $ 225.5 million for these purposes.

Today, the US DoD is working on the following main combat hypersound programs: Navy – Conventional Prompt Strike (CPS) and Offensive Anti-Surface Warfare Increment 2 (OASuW Inc 2; also often – Hypersonic Air-Launched OASuW or HALO); Air Force – AGM-183 Air-Launched Rapid Response Weapon (ARRW) and Hypersonic Attack Cruise Missile (HACM); CB – Long-Range Hypersonic Weapon (LRHW); DARPA Agency – Tactical Boost Glide (TBG), Operational Fires (OpFires) and Hypersonic Air-breathing Weapon Concept follow-on (MoHAWC).

The result should be the creation of existing prototypes, and the best of them will go into series and into service, which will increase the capabilities of the US Armed Forces to defeat important enemy targets (mobile missile launchers, air defense and missile defense systems, etc.). However, some experts say that the Pentagon has yet to develop a concept for the use of hypersonic weapons.

Influence is important, but poorly predictableWhat conclusion do the authors of the report make?

The high degree of influence of advanced military technologies and weapons on the outcome of armed confrontation on the battlefields of the wars of the future is recognized: "Many new technologies ... can potentially influence the nature of the war of the future. In particular, advances in areas such as artificial intelligence, big data analysis and deadly autonomous systems will reduce or eliminate the need to use a human as an operator.… Low-cost unmanned aerial vehicles can change the balance between quality, which the US Armed Forces have traditionally relied on, and quantity, as well as between offensive and defensive. So, swarms of unmanned vehicles can overwhelm defense systems, providing a greater advantage to the attacker, while directed energy weapons, which provide an inexpensive means of neutralizing such attacks, can benefit the defender."

The complexity of predicting such an impact is also pointed out: "The impact of new technologies on the course of hostilities and strategic stability is difficult – if not impossible – to predict, since it will depend on many factors, including the pace of technological progress both in the United States and in rival states; the way new technologies are integrated into existing armed forces and operational concepts; the interaction between new technologies and the extent to which national policies and international law allow or hinder their development, integration and use."

The final summary says that in the coming decades, new technologies can repeatedly change the balance of attack and defense, which may have "unforeseen consequences for the conduct of hostilities and strategic stability."

In 2014, the Pentagon put into effect the Third Offset Strategy and the Defense Innovation Initiative, the goal of which is to preserve America's military–technical superiority in the current century. The United States intends to achieve this by making extensive use of the latest technologies and intelligent solutions, as well as improving the concepts of the use of forces and means at all levels.

The importance of new military technologies capable of changing the very nature of war is noted in the US National Defense Strategy (NDF) of 2018, as well as in the recently approved NDF of 2022. We hope that the Russian leadership has made a similar conclusion about the role of new military technologies in the war of the future.

Vladimir Shcherbakov

Deputy Executive Editor of HBOVladimir Leonidovich Shcherbakov is a military expert, historian, and writer.