Community Content
This article has been written and edited by the research team. It has not passed through the quality control procedures applied to content produced by the Research Outreach team.
October 10, 2023

Artificial Intelligence as Dual-Use Technology

This paper discusses dual-use technologies (DUT), technologies that have potential to be used for both peaceful and military aims, from a global perspective, especially with regard to Artificial Intelligence (AI). Strategic research and development (R&D) of DUT has its roots in the Defense Advanced Research Projects Agency (DARPA) in the United States (US). DARPA was established under the US DoD's national strategy to overcome the, so-called Sputnik shock and its success has led to the realization of the current information society. Since then, major countries around the world have established DARPA-like R&D organizations and are operating on R&D of DUT. AI is one of key technologies for DUT; thus, its concepts and characteristics are discussed. On the other hand, DUT is an aggregation of strategic cutting-edge technology for each country, which is important for national security and defence. In particular, since AI is in a form of software, outcomes can easily leak or get stolen through the Internet. This issue is known as 'dual-use dilemma' and precautions must be taken against it. Therefore, countries, organisations, and universities are taking and/or considering various counter measures. At the same time, the Japanese academic community is currently in denial about DUT-policy based on its unique background. The author hopes that this chapter will be helpful in future works on this topic.


Dual-use technology (DUT) is a technology that has the potential to be used for both peaceful and military purposes, and as such can be a crucial asset for a nation. One of the key DUTs is artificial intelligence (AI), as this can improve the intelligence of any machine or system to a far higher level. Research and development (R&D) of AI DUT has become a key issue for nations around the world.

The author considers the issue on R&D of DUT from a global perspective, based on a review of the concept of AI and its recent trends. It is widely recognised that DARPA (Defense Advanced Research Project Agency) in the US has led the charge in R&D of AI, and therefore the issues of R&D of DUT regarding DARPA are discussed as the key topic.

DARPA was established in 1958 to strengthen national power and surpass the so-called ‘Sputnik Shock’ which occurred in 1957. Establishing this organisation also had the intention of obtaining world leadership in both basic research of science and technology (S&T) and military technology. DARPA made a big success in the history in DUT R&D, and created the so-called ‘DARPA model’, one which consists of close collaboration between academia, industry, and military to achieve the innovative goals proposed by DARPA.

The US military is one of the world leaders in terms of research and military power, some of which may relate to how modern information society has been built by a variety of advanced IT technologies such as Internet and IoT. Many major countries have established DARPA-like organisations, or are in the planning stages for such organisations.

Basic models of AI

AI was proposed as a new scientific area in computer science by a group of young computer scientists interested in the idea of ‘making machines (computers) intelligent alike human’ at the Dartmouth Workshop held in 1956 in US.

AI is generally classified into three basic models as shown in the image above. Knowledge-based AI was first AI model proposed in the 1970s and was mainly used for Expert Systems. The introduction of knowledge-based AI generated the first AI boom. Agent-based AI was proposed in early 1990s, and was mainly used for autonomous systems. Deep Learning was the newest AI model that has obtained dramatic attention with a computer Go program AlphaGo in 2016, and generated another AI boom.

Realising human-like intelligence with computers does not necessarily require the same intelligence mechanism as in humans. Recently, AI has been expanded to include the realisation of functions that are generally considered to be ‘smart’. DUT is a perfect example of a ‘smart system’.

DARPA-like organisations

DARPA has changed the paradigm of DUT R&D. Prior to establishing DARPA, cutting-edge technology (CET) used to be developed by the military for use in wartime. After the technology had fulfilled its purpose during the war, it would then be used in more peaceful pursuits. Since the establishment of DARPA, DUT R&D has been achieved by collaboration during peacetime, instead of during conflicts.

DARPA has also proposed ‘high-risk, high-return’ goals with big R&D funds and considerable guidance. Research universities and industry have benefited from this paradigm. The US has since obtained the world leadership in academia, industry, and security with reasonable costs, further inspiring other major counties to establish DARPA-like organisations.

The UK is planning to establish an Advanced Research and Invention Agency (ARIA) to obtain a global superpower as an independent organisation. China established the State Administration of Science, Technology and Industry for National Defense (SASTIND), under which are seven universities called ‘national defence universities’. India established the Defence Research and Development Organisation which has its own in-house R&D laboratory system. In Germany, The Ministry of Education has a policy of developing ‘Germany’s own AI’ to gain global leadership, and military applications seems to be included based as a necessity. Japan’s academia is negative on DUT R&D. SCJ, the representative body of the academic sector, has announced the statement of strong negation against DUT for military technology at university.

Dual-use dilemma

There are international concerns about the ‘misuse’ of DUT R&D outcomes which include confidential information and knowledge on national security, and must be strongly managed by means of suitable measures. This kind of problems are referred to as ‘dual-use dilemma’, that is the sensitive issue at global scale. The dilemma is basically the issue of ethics.

The measures include export controls, preventing cyberattacks, and the establishment and dissemination of guidelines for international collaborative research at universities. Since AI technology is digital, there is the fear that sensitive data will be exposed to risks from online hacking. As such, R&D of preventive measures is necessary.



Ueno, H, (2023) Artificial Intelligence as Dual-Use Technology. In Loannis, K, Hatzilygeroudis, et al, (ed) Fusion of Machine Learning Paradigms, Springer, Cham.

Written By

Haruki UENO
National Institute of Informatics

Contact Details


Want to read more articles like this?

Sign up to our mailing list and read about the topics that matter to you the most.
Sign Up!

Leave a Reply

Your email address will not be published. Required fields are marked *