ROLE OF MILITARY ARTIFICIAL INTELLIGENCE

IN THE EVOLUTION OF CONFLICTS

Lieutenant-general (ret.) Professor Cristea DUMITRU, Ph.D^

(Academy of Romanian Scientists, 3 Ilfov Street, 050044, Bucharest,

Romania, email: secretariat@aosr.ro)

Abstract: Artificial intelligence may accelerate the pace of war, but security

architecture must also accelerate and deepen at the same pace, building a new

generation of capabilities.

The real challenge in military artificial intelligence (AI) is to maintain

human accountability and political wisdom while maintaining technological

superiority as we move toward algorithmic dominance. Therefore, every technical,

legal, and ethical step taken today will shape not only the nature of war but also the

future of the international order.

Keywords: military artificial intelligence, security architecture, AI norms,

AI in military operations, human control in military AI, hyper-war.

DOI

10.56082/annalsarscimilit.2026.1.19

Military experts believe that artificial intelligence (AI) can accelerate

the pace of war and determine combat success, but the security architecture

must accelerate, adapt, and deepen at the same pace, building and developing

a new generation of capabilities. If this does not happen, the fragility/entropy

of strategic stability will increase within the millisecond decision cycles of

next-generation AI-powered warfare technology.

Periodically, those responsible for AI in the military meet in summits

focused on the responsible use of artificial intelligence in the military. Thus,

such meetings were held in The Hague in 2023, Seoul in 2024 and in February

2026 in Spain. After this year's summit, which was attended by experts from

85 countries, a declaration was developed that includes 20 key principles.

Thus, the principles assumed in the aforementioned declaration include

maintaining human responsibility for AI-based weapons, establishing clear

chains of command and control, exchanging information on national

oversight mechanisms, focusing on risk assessment, training and testing

processes for military personnel. At the same time, this approach is

considered too cautious by some states, especially those states where

investments in AI are paramount. This excessive caution and regulation is

considered to induce critical delays in the impetuous development of AI and

quantum technologies, and any delay or loss of initiative in this extremely

^Entitled member of the Academy of Romanian Scientists, Military Sciences Section, email:

cristea.dumitru@force1.ro.

19

ROLE OF MILITARY ARTIFICIAL INTELLIGENCE

IN THE EVOLUTION OF CONFLICTS

dynamic field, in military terms, can constitute a catastrophic deficit. The

reality is that we are living in the beginning of a new era of humanity, the era

of AI, an era of intelligent robots, quantum computing, technologies that

change the social order based on work. This colossal transformation generates

seismic social imbalances, which almost inevitably lead to the triggering of

both internal and international conflicts, difficult to control and stop. The

anguish of this perspective induces the solution of cautious regulations,

although these would have no effective effect, while precisely the

acceleration of AI development and reaching a threshold of general

superintelligence would provide systemic solutions to social problems and

eliminate the root causes of world conflicts.

It is clear that decades of ignoring the potential of technological

advances by the political and economic spheres have generated surprise, a

total lack of preparation and consequently a setback in the wide-scale

adoption of AI and the tendency towards overregulation, including in the

military. The risk is that, in the military, the actor who breaks the rules can

create an insurmountable advantage in a possible conflict, if he achieves AI

dominance. At the same time, once AI dominance is ensured, there is no

chance of the adversary coming back, and the loss is final and cataclysmic.

Thus, future conflicts will be conflicts for the dominance of algorithms.

At the REAIM (Responsibility for AI in the Military) summits in 2023

and 2024, many countries, coordinated by the US, adopted a more sensible

framework of principles. This year, however, the fact that two leading

countries in the field of AI technology, the USA and China, did not sign the

declaration and that the number of countries that ratified it was limited (35),

demonstrated how limited the production and sharing of a set of

internationally accepted norms and principles is. Unfortunately, so far, a

common denominator has not yet been established in the international system

regarding AI in the military.

This situation, in addition to the profound transformation triggered by

AI technologies in the operational environment, can be considered a threat to

security.

Military Artificial Intelligence in the Context of Security Threats

Currently, the use of AI in the military context covers a wide range of

areas, from command and control and decision support to electronic and cyber

warfare, missile defense and unmanned aerial, land and naval vehicles.

In numerous scenarios, which exceed human cognitive capacity or

require rapid analysis of large and complex data sets, AI functions as a key

element. The technological processes already place almost all the necessities

of the modern battlefield in this category. In other words, AI already takes

warfare to a dimension that goes beyond human perception and reaction. It is

already a reality that, in this new type of warfare, defined in the specialized

20

Lieutenant-general (ret.) Professor Cristea DUMITRU, Ph.D

literature as “HYPER-WAR”, the decision-making and execution cycle can

be completed in milliseconds.

It is clear that at this point, it becomes necessary to consider human

control, from a much broader perspective and necessity than the simple

“finger that pulls the trigger”.

It is therefore necessary that the first step in this direction is to design

human control as a significant and effective component in all stages of the

processes of using AI. Such a structure would ensure that the ethical and legal

responsibility for the use of artificial intelligence remains entirely under

human control, responsibility and competence. At the same time, the

vulnerabilities of AI technologies must also be taken into account, the greatest

of which being strong electromagnetic pulses, vulnerabilities that could

quickly eliminate the advantages of an ultra-technological army.

The second step would be to develop filters and barriers to prevent

systems and processes directly or indirectly managed by AI from “negatively

feeding” each other, interacting and escalating conflict. Such a multi-level

filtering structure could prevent false alarms, “AI hallucinations”, lack of

awareness or unwanted engagements.

The third crucial step in managing internal security risks in military

AI could be validating the security and integrity of the algorithms and the

datasets they feed. Preventing contamination of datasets, identifying

conditions and times when AI algorithms cannot make fully autonomous

decisions, as well as certifying and monitoring the algorithms and the

hardware they run on, are important measures in this context.

However, it is important to highlight that the lack of a minimum global

consensus on military AI makes it impossible to adopt structured ethical, legal

and normative measures in this area. We are witnessing today the worsening

of the existing international strategic trust deficit and even more so by the

attractiveness of the opportunities and capabilities offered by military

artificial intelligence; the concerns expressed from time to time by various

lobby groups and states under headings such as: human rights, confidentiality

and protection of personal data become weak in the face of the increasing

effectiveness of artificial intelligence in military operations.

Regulation of Artificial Intelligence. Limits and Requirements

At this point, we must ask ourselves how we can determine whether

an AI software or algorithm is safe or “human”. Among AI experts, it is

proposed that a structure similar to the International Atomic Energy Agency

(IAEA), which oversees nuclear facilities, be established for military artificial

intelligence. But it must be emphasized that the codes that make up the

algorithm do not leave traces like radioactive elements and tracing the source

of the code is much more difficult than supervising the establishment and

operation of a nuclear facility. However, as a solution, it is proposed to

21

ROLE OF MILITARY ARTIFICIAL INTELLIGENCE

IN THE EVOLUTION OF CONFLICTS

establish “algorithmic control rooms” in which countries would test and

verify artificial intelligence systems without revealing their fundamental

architecture and confidential components. Thus, the certification resulting

from tests carried out in these simulation environments could set a standard

in the commercial and military fields.

It is therefore important to emphasize that the main issue is not

whether AI will be used in the battlespace, but how, by whom and within

what limits it will be used. At the same time, it is clearly shown that the

development of norms in the field of military artificial intelligence cannot be

considered independent of the competition between great powers. It is

necessary that technology leaders such as the US and China intensify their

participation in the development of norms with a view to creating a globally

accepted ethical framework in the field of generative AI.

For example, China has stated that its long-planned military

modernization will proceed in three distinct but overlapping phases:

mechanization, i.e., the adoption of modern machinery and equipment;

informatization/digitalization, i.e., the integration of advanced information

technologies and cyber networks to connect military platforms and enable

real-time information exchange; intelligence, the application of artificial

intelligence to automate operations, support decision-making, and control

advanced weapons. Today, China is placing greater emphasis on integrating

AI into its military by prototyping AI capabilities that can pilot unmanned

combat vehicles, detect and respond to cyberattacks, track naval vessels,

submarines, and identify and strike targets on land, at sea, and in space.

The Chinese military is also developing systems that ingest, analyze,

and augment massive amounts of data to improve tactical and strategic

decision-making, as well as tools that create deepfake images and videos for

disinformation campaigns. Accordingly, Beijing believes that the military

that better develops and adopts artificial intelligence and other emerging

technologies will gain a major advantage in future wars, especially those AI

technologies that support and accelerate decision-making. The Chinese

military’s AI experiments also extend to cyber and information operations,

developing AI tools to automate the detection of intrusions into its computer

networks, to increase the resilience of military communications, and to

enhance its cyber operations. Meanwhile, Chinese officers and soldiers are

using AI systems to simulate virtual battlefields and model the behavior of

competitors, which improves their training for future conflicts.

The Chinese military is increasingly applying AI to diminish U.S.

advantages in space and maritime. It is openly pursuing satellite-targeting

algorithms as well as new anti-satellite weapons, some of which involve small

robots that can capture and disable an adversary’s space platforms. China

aims to build a military that leverages advanced technologies to learn, adopt,

and make decisions quickly and more accurately across all operational

22

Lieutenant-general (ret.) Professor Cristea DUMITRU, Ph.D

domains. China’s path will certainly not be smooth. The Chinese military still

faces obstacles in integrating artificial intelligence across its entire force and

operations. Moreover, training AI systems requires a large amount of data

that is not readily available on the internet, such as classified images of

military platforms or the electromagnetic signatures of various radars and

weapons.

Unless technological leaders like the U.S. and China participate in the

discussions, efforts to create an accepted global ethical framework will

inevitably remain limited.

At the national level, institutionalizing measures such as meaningful

human oversight, algorithmic security, data integrity, and multi-layered

filtering mechanisms is an urgent need. This is because a significant portion

of the risks posed by military AI come not from intentional misuse, but from

design flaws, inaccurate data, unpredictable interactions, and the pressure for

speed.

 



Artificial intelligence may accelerate warfare, but security

architecture must also accelerate and deepen at the same pace, building a new

generation of capabilities.

The real challenge in military artificial intelligence (AI) is to maintain

human accountability and political wisdom while maintaining technological

superiority. Therefore, every technical, legal, and ethical step taken today will

shape not only the nature of warfare but also the future of the international

order.

BIBLIOGRAPHY

HERMAN A., „China's Brave New World Of AI," Forbes, august 30, 2018,

available

at

https://www.forbes.com/sites/arthurherman/-

2018/08/30/chinas-brave-new-world-of-ai/ - 3a7918bf28e9;

HOROWITZ M. „Artificial Intelligence, International Competition, and the

Balance of Power”, Texas National Security Review, available at

https://tnsr.org/2018/05/artificial-intelligence-international-

competition-and-the-balance-of-power/;

HOROWITZ M., „The promise and peril of military applications of artificial

intelligence”, Bulletin of the Atomic Scientists, available at

https://thebulletin.org/2018/04/the-promise-and-peril-of-

military-applications-of-artificial-intelligence/;

23

ROLE OF MILITARY ARTIFICIAL INTELLIGENCE

IN THE EVOLUTION OF CONFLICTS

TALBERT E. „Capcana triajului. Când viteza AI înlocuiește judecata la

comandă”, january 2026;

TUCKER P., „How NATO's Transformation Chief is Pushing the Alliance to

Keep

Up

in

AI",

Defense

One,

available

at

https://www.defenseone.com/technology/2018/05/how

nato's

transformation chief is pushing the alliance to keep up in

ai/148301/.

AI Principles: Recomandări privind utilizarea etică a inteligenței artificiale de

către Departamentul Apărării" Defense Innovation Board,

October 31, 2019, available at https://media.defense.gov/2019/-

Oct/31/2002204458/-1/-1/0/DIB_AI_PRINCIPLES_-

PRIMARY_DOCUMENT.PDF;

„Despre noi", NATO Center of Excellence for Strategic Communication,

available at https://www.stratcomcoe.org/about-us;

„Federated Mission Networking," NATO Allied Command Transformation,

available at https://www.act.nato.int/activities/fmn;

Electronic sources

www.digi24.ro;

www.antena3.ro.

24