The aspects of the battlefield unfolding across the globe in the 2020s are showing
patterns that surpass the human history of over 2,000 years, even exceeding the
devastation of World War I and II, which resulted in the worst human casualties.
In this...
The aspects of the battlefield unfolding across the globe in the 2020s are showing
patterns that surpass the human history of over 2,000 years, even exceeding the
devastation of World War I and II, which resulted in the worst human casualties.
In this era, where the importance of humanity is magnified due to an
unprecedented demographic cliff in developed nations worldwide, the emergence of
Artificial Intelligence (AI) and Manned-Unmanned Teaming (MUM-T) combat
systems, designed to reduce human casualties in combat, is fundamentally
transforming the paradigm of the battlefield.
The future warfare, based on hyper-connected and hyper-intelligent technologies,
no longer relies solely on the physical performance of hardware-centric weapon
systems and large platforms with human operators. Instead, it is evolving into
'Intelligent Weapon Systems and Intelligent Defense Systems,' where software and
data are central. However, for the Republic of Korea to acquire such weapon
systems, there are still many hurdles to overcome. The current system remains
rooted in the industrial era, creating a deep gap between the rapidly advancing
pace of technology and the innovation required by the national defense sector. The
current defense acquisition system of the Republic of Korea contains structural
limitations that act as significant obstacles to the introduction of new technologies,
such as AI-based MUM-T systems.
The purpose of this study is to overcome the gap between this technological
reality and institutional limitations, and to improve the institutionalization plan for
the rapid adoption of AI-based MUM-T systems. Accordingly, by proposing a new
legal and institutional framework suitable for the AI era, we aim to present a
blueprint necessary to transition the defense acquisition system itself from a
'hardware-centric, closed acquisition system' to a 'software-centric, open innovation
ecosystem,' going beyond simple deregulation. To achieve this, the following
problems are presented, and solutions are sought
First, the issue of rigid systems can be examined. The current Defense Acquisition
Program Act is optimized for the acquisition procedures of conventional large
weapon systems, necessitating an overly protracted and staged process from
requirements determination to final deployment. This makes it difficult to reflect the
characteristics of AI and modern technology, which evolve at real-time speeds,
leading to the problem of 'technological obsolescence,' where the technology is
already outdated by the time it is deployed.
Second, there are legal barriers hindering the flow of data and technology. The
development of AI technology presupposes the effective influx of high-quality data
and open civilian innovation into the defense sector. However, the strict application
of the Military Secret Protection Act makes the utilization of military-generated
data difficult, and the Defense Industry Technology Protection Act tends to restrict
open technological exchange with the private sector. This isolates the defense
sector from the open and innovative civilian ecosystem, exacerbating the isolation
of data development in the defense field.
Third, complex problems arise due to inefficient organizational operation. Despite
AI technology being a field driven by specialized expertise, the personnel
management systems of the organizations responsible for defense acquisition
management and the defense industry regulatory bodies remain in the past. The
'trap of rotational assignment,' where personnel change posts every 1 to 2 years,
makes it difficult to accumulate technical expertise, hinders business continuity, and
the rigid organizational culture that sees 'failure as immediate disciplinary action'
fundamentally stifles new attempts and the spirit of challenge, blocking innovation.
Through this research, we contemplate not merely developing or improving the
defense acquisition management system for the era of transition to AI-based
MUM-T systems, but a comprehensive direction for improving laws, systems,
organizations, and human networks. The results must pave a new path for the rapid
realization of AI-based MUM-T system acquisition.
First, we propose the establishment of a special path or the improvement of the
system for the rapid acquisition of innovative technologies specialized for the
defense sector. This is to overcome the limitations of the current defense
acquisition system and to legally guarantee a flexible procedure with the principles
of speed and flexibility, allowing field units or researchers to first discover the
technological potential and propose requirements through demands raised in the
combat field.
Second, instead of the existing linear development based on large hardware, the
'Agile' method is introduced, which allows for 'fast failure, fast learning,' iteratively
improving technology in short cycles.
Third, once the military utility of a new technology is proven through rapid
deployment, it is quickly put into service without complex procedures, thereby
simultaneously solving the 'valley of death' where technology is shelved and the
'regulatory barrier' that slows down development due to regulations.
Fourth, innovation obstacles must be removed by opening up the 'blocked paths'
created by various regulations and laws. No matter how fast a path exists, the flow
of data, technology, and radio waves is blocked. Therefore, passageways must be
secured through the following legal revisions
Amend the Military Secret Protection Act to introduce the concept of Defense Data
with specialized security classifications that can be used restrictedly in the civilian
sector for the utilization of military data, which is key to AI development. Special
provisions should be established to allow data sharing and utilization with private
companies while maintaining security.
Amend the Defense Industry Technology Protection Act to establish provisions that
exempt basic research for free technological exchange with the private sector, and
clarify the scope of technology protection to expand the range of open
collaboration.
Amend the Radio Waves Act to designate a 'Frequency Regulatory Sandbox for
Defense Innovation' to relax frequency regulations essential for the development
and operation of AI-based MUM-T systems, allowing flexible frequency usage only
for research and demonstration purposes to alleviate the scope of restrictions.
Fifth, and finally, concentrate the best efforts on organizational and cultural
innovation and the nurturing of experts. Since the agents who will drive innovation
through the introduction of new systems are ultimately people and organizations,
innovation in personnel and organizational management must be pursued in parallel.
Firstly, an organization should be established to directly infuse the innovative
capacity of the private sector and take exclusive charge of rapid acquisition
programs for a major overhaul. This organization should be granted the authority
for flexible budget execution and rapid decision-making, responsible for discovering
private technology and managing small-scale demonstration projects. Within this
framework, to foster and utilize specialized personnel, the 'trap of rotational
assignment' where public officials continuously change posts according to their
years of service must be broken. A new specialization focusing on AI-based
MUM-T systems should be established, and legal grounds should be prepared to
assign specific personnel equipped with specialized knowledge and skills for
long-term positions and deepen their expertise. Furthermore, systems such as an
'Advanced Science and Technology Reserve Force' should be expanded to utilize
civilian AI experts in military research. Research and challenges undertaken by
these individuals should not be hesitated upon, and project failures, unless caused
by intent or gross negligence, should be considered as 'opportunities for new
experience,' with a system established to share the lessons learned. This is
absolutely essential for breaking away from rigid bureaucracy and promoting the
spirit of challenge.
The system improvement plans proposed in this study are not individual policies
but an organic package in which three axes—procedural innovation, barrier removal,
and personnel and organizational innovation—must interlock and be pursued
simultaneously to bring about meaningful change.
AI-based MUM-T systems are no longer a story of the future but a reality that
fundamentally changes the paradigm of national security. If laws and systems fail to
keep pace with this technological progress, it will lead to the erosion of national
security capabilities, beyond just the loss of technological superiority. Therefore,
the proposed improvements in this study are more than just an efficient
transformation of the acquisition process; they are the minimum legal and
institutional preparation necessary to secure the nation's survival against the
security threats of the coming AI era.