Denmark is one of the most digitized societies in the world. The Danish defence and security sector is known for its technological advancements and commitment to innovation. Denmark is currently on the verge of pursuing defence AI through a range of emergent projects and initiatives. Moreover, the Danish armed forces are packed with advanced military technology that has the potential to collect large amounts of raw data with the help of radars and sensors.

The problem is, however, that data is not always used properly because it rarely leaves the frontline. Danish Defence therefore needs to adopt AI as a part of the development of a much stronger digital backbone to process the increasing amounts of information and share it across platforms, units, and domains. This is a prerequisite for integrating forces into synchronous, multi-domain operations; a goal Denmark is currently pursuing as a member of the NATO alliance.

Commenting on this strategic thinking, Danish Chief of Defence, Flemming Lentfer, has stated how he sees “a need for servers before fighter jets, ships and tanks,” (Krog 2021) a viewpoint which is supported by several voices in the Danish defence and security environment. For instance, in the report Danish Security and Defence Towards 2035, the government’s security policy analysis group notes how the “rapid technological development will require a significant technological boost to the Danish Armed Forces if it is to remain a relevant partner for our strongest allies” (The security policy analysis group 2022: 5). However, the question of how exactly this technological boost should be accomplished and implemented into the organisation remains uncertain and uncoordinated.

The path to adopt defence AI in Denmark is challenged by fragmentation and lack of governance and strategic vision. Whilst testing and evaluation of AI is starting to sprout from various branches of the organisation, these initial efforts are mostly isolated and uncoordinated. The potential benefits in this bottom-up approach—such as explorative and experimental initiatives and a higher willingness to take risks—does not compensate for the lack of strategic direction.

An imperative task therefore remains in formulating a Danish strategy for defence AI with specific focus on where in the organisation and how AI should contribute to the “technological boost” of the Danish Defence. This includes a more precise definition of AI and what it should do for the military organisation, operationally as well as in the support functions. Second, the recent paradigm shift, changed approach and establishment of public-private partnerships within the Danish defence industry and academia should be utilized to strengthen development of AI and AI-enabled software solutions in Denmark.

Finally, the need for testing, training, and evaluation, as well as education of future defence AI specialists should be included and considered within the framework of existing and new military education programmes in Denmark. This is crucial to recruit and retain the right digital skills and competences needed for future Danish AI-enabled multi-domain operations and to build and sustain a modern data-driven defence organisation.

1 Thinking About Defence AI

As a small nation, Denmark is in the early stage of adopting defence artificial intelligence (AI). Recognizing the increasingly central role AI plays in the global technology competition and on current battlefields, it has nevertheless become clear that Danish Defence needs a major technological boost to remain relevant for its allies (The security policy analysis group 2022: 5).

The intensified technological competition not only reshapes security policies for smaller states like Denmark but also influences the international framework for Danish defence and security policy. Denmark has a unique opportunity to contribute to and benefit from the evolving landscape. However, without a clear technological reform agenda, the Danish Defense risks disconnection from close allies (Breitenbauch and Liebetrau 2021: 9).

Accordingly, the Danish Ministry of Defence is about to launch a range of strategic initiatives on AI, data, and digitalization. These include setting a direction for a digital transformation of the organisation and establishing strong and robust connections between operational and supporting processes, AI and data analysis, IT support, and digital competence development among management and employees.

Supposedly, the upcoming strategic initiatives, including a strategy for digital transformation and a data strategy, will address a major challenge in relation to AI in Danish Defence: a growing gap between military and digital knowledge due to a general lack of AI literacy and understanding of how data can generate value for the organisation. These challenges remain significant barriers for adopting defence AI, which makes conscription and recruitment of IT-specialists, technical translators, and experts in data science a still more demanding task. Thus, the challenge of cultivating a digital mindset for exploiting the possibilities of AI is becoming a top priority for Danish Defence (Bollmann and Jacobsen 2023).

In short, the people in the organisation should learn to see data as a strategic asset that can provide tactical advantages and strategic autonomy for a small state like Denmark. It is thus key to prioritize education and enhance the understanding of technology, digital literacy and trust in data, AI, and autonomous systems. But above all, it remains pertinent to consider the broader imagination about defence AI among employees and managers as well as the public debate among civil actors in academia and industry.

1.1 Imagining AI in Denmark

In Denmark, the common perception, discourses and imaginaries surrounding defence AI have been formed by popular culture and science fiction scenarios envisioning a future world of killer robots and super computers (like Skynet in Terminator or HAL in The Space Odyssey) turning against their human creators. Although arguably far from reality, such dystopic visions are still recurring themes in Danish media and public debate about future military AI in Denmark. For instance, the fictional video Slaughterbots created by the Campaign to Stop Killer Robots in 2018 is frequently featured in the Danish public debate and in campaigns warning against autonomous weapons. Another example: in a critical campaign called Autonome Våben, Slaughterbots features together with the armed drone called Lanius, developed by Israeli Elbit Systems. According to the featured promotion video for Lanius, the drone uses artificial intelligence and camera sensors to track down and kill enemies (Autonome Våben 2023).

The Danish debate on defence AI is further accelerated by the wave of critical opinions about AI in relation to ChatGPT such as the critique stated by the Future of Life Institute in March 2023 (Future of Life 2023). In an open letter, prominent voices such as Elon Musk called on all AI labs to immediately pause the training of AI systems more powerful than GPT-4 for at least 6 months. Inspired by this critique, Danish scientists and researchers made a similar statement in June 2023. In an open letter to the Danish Minister for Foreign Affairs, Lars Løkke Rasmussen, the 20 researchers called for the Danish government to support an international ban on autonomous weapons systems. According to the researchers behind the letter, the Danish government should therefore develop a national policy for defence AI with the goal of a legally binding international agreement that “prohibits the use of autonomous weapons which is not under meaningful human control or has humans as targets” (Nedergaard et al. 2023).

The response from the Danish Foreign Minister was a rejection of any specific Danish policy, ban or legislation against autonomous weapons. However, the minister stated how he sees a “need for an international framework” and that Denmark together with other European countries supports a “two-step approach” to autonomous weapons where regulations according to international humanitarian law is central (Pilgaard 2023: 5).

The open letter from the Danish researchers addresses a sensitive issue in the official Danish approach to defence AI: a general lack of governance and political guidelines. The vague response from the Minister for Foreign Affairs is therefore symptomatic of the official Danish view offensive aspects of defence AI: to outsource such matters to international fora such as the United Nations. Moreover, the AI debate in Denmark discloses a lack of conceptual clarity in the Danish political and public discourse about defence AI (and AI in general). This lack is mainly due a gap between the national level of ambition and the fact that no political or strategic guidelines have so far been formulated on defence AI in Denmark.

1.2 Denmark as AI “Front-Runner”

In 2019, the Danish National Strategy for Artificial Intelligence was launched with the ambition that Denmark should “be a front-runner in responsible development and use of AI,” (National Strategy for Artificial Intelligence 2019: 5). The strategy defines AI in highly generic terms based on principles from the OECD and the European Commission, as:

systems based on algorithms (mathematical formulae) that, by analyzing and identifying patterns in data, can identify the most appropriate solution. Most of these systems perform specific tasks in limited areas, e.g., control, prediction, and guidance. The technology can be designed to adapt its behavior by observing how the environment is influenced by previous actions (National Strategy for Artificial Intelligence 2019: 6).

Defence and security are entirely absent. Thus, the text is not at all specific about how the military should implement and use AI while being exposed to the battlefield’s many dilemmas. Instead, the strategy partly builds on ethical principles framed in the Declaration on AI in the Nordic-Baltic Region signed by the Ministers responsible for digital development from Denmark, Estonia, Finland, the Faroe Islands, Iceland, Latvia, Lithuania, Norway, Sweden, and the Åland Islands. In this declaration the countries agreed to collaborate to “develop and promote the use of artificial intelligence to serve humans” (Nordic Council of Ministers 2018: 1). Moreover, the Danish AI strategy supports the declaration objectives for making the Nordic-Baltic region a digital leader and embraces the European Commission Communication on “Artificial Intelligence for Europe” and the declaration of 24 EU Member States and Norway on “Cooperation on Artificial Intelligence.”

The strategy focuses mostly on the possibilities for AI to contribute to better public-sector services and growth in the business community. It highlights how a safe and responsible use of AI should follow six ethical principles which are to be incorporated in the development and use of artificial intelligence to “secure respect for individuals and their rights, and for democracy” (National Strategy for Artificial Intelligence 2019: 8). The six ethical principles are:

  1. 1.

    Self-determination

  2. 2.

    Dignity

  3. 3.

    Responsibility

  4. 4.

    Explainability

  5. 5.

    Equality and justice

  6. 6.

    Development

Some of the goals pertaining to the principle Dignity include that “AI should not cause injury, it should support due process and it should not unjustifiably place people in a worse position.” Moreover, “AI should not be used to infringe fundamental human rights” (National Strategy for Artificial Intelligence 2019: 28).

If these goals are understood literally, it seems unlikely that AI will be used in relation to power projection by the Danish military anytime soon—as long as AI-enabled warfighting power requires a legal and ethical mandate to actively and effectively cause harm, or threaten to cause harm, to humans.

Although not taking any stance on this issue, the Danish ambitions for a responsible and explainable development and use of defence AI are aligned with international norms and standards such as the principles for safe and responsible use of defence AI in the NATO AI strategy (NATO Artificial Intelligence Strategy 2021). Denmark adheres to the common notion that humans must remain in full and constant control and that AI must be hierarchically subordinated to humans. But unlike other countries, Denmark has not yet formulated any specific (ethical, juridical, or political) guidelines for the development and use of defence AI.

This absence of political or strategic guidelines has been critiqued by military commentators and legal experts. For instance, it has been questioned how the Danish Airforce will make proper use of the sophisticated AI technology built into newly procured platforms, such as the F-35 fighter jets, if it is to comply with the ethical principles formulated in the AI Strategy (Jarlner 2021). This critique exposes the paradox inherent in the Danish defence AI imaginary and the double ambition to be an AI frontrunner while simultaneously restricting these ambitions to ethics and security policies of defence AI.

Despite the absence of specific political guidelines for military use of AI, digitalization and datafication are rising priorities for the Danish Armed Forces. Danish Chief of Defence, Flemming Lentfer, in an interview noted that he sees “a need for servers before fighter jets, ships and tanks,” (Krog 2021), thereby advising the politicians to invest accordingly. The statement provoked critical reactions among military analysts and commentators, arguing that data centres and AI do not compensate for the need for conventional military capacities and hardware (Nielsen 2021). The message from critics appears to question investing heavily in headquarters with splendid overviews and situational awareness if the military on the field lacks the necessary equipment to act.

1.3 A Shift in Strategic Culture

It seems as if the statement from the Chief of Defence marks a shift in strategic culture and procurement practice for the Danish Defence. The statement not only comes from the highest ranking military officer of the Danish Armed Forces, but also seems to be aligned political initiatives coming from the Ministry of Defence and the implementation of new and more agile procurement processes and practices in the Danish Defence Acquisition and Logistics Organisation (DALO).

Traditionally, Denmark has invested its military capabilities in weapons systems such as artillery and tanks, especially from US, with which Denmark shares a strong bond. Historically, the strategic culture and thinking in Denmark has been tied to the military principles and priorities of NATO and the foreign policy of the USA. When Denmark decided in 2001 to participate in the war in Afghanistan just three months after the terrorist attacks on 11 September 2001, it was an act of solidarity with the United States (Rasmussen 2005: 67). Already in the 1990s, the Danish military engagement in Kosovo marked a new form of military activism that broke with the more “defencist” passivity that marked the Danish security politics and discourse in the late 1970s and 1980s. This also included a new focus on European integration and globalization, giving a new interpretation to military power in Denmark.

The transformation of the Danish strategic thinking towards a culture of activism and cosmopolitism has naturally reached a new peak with the Russian invasion of Ukraine and the following Danish donation politics of military equipment. This not only includes tanks and F-16 fighter jets, but also capacities for intelligence, surveillance and reconnaissance. This includes drones equipped with AI-enabled sensor and software systems produced and delivered by Danish companies such as Nordic Wing and Skywatch in 2022 and 2023 (Dall and Lindegaard 2023). The latter’s promotion efforts for their RQ-35 Heidrun drones were assisted by former Star Wars actor Mark Hamill (Luke Skywalker) as part of a project crowd-funding drones for Ukraine (Skywatch 2023). This emphasizes the role of media, public perception and the broader cultural drone and AI imaginaries in promoting new military technologies for the battlefield.

The threat from Russia has changed the geopolitical situation and increased the salience of the Baltic and Arctic regions for Denmark’s security. This situation helps foster a renaissance of Denmark’s maritime tradition through a reorientation towards maritime security and the protection of Danish territory (including Denmark, Greenland, and the Faroe Islands) and its immediate surroundings. Moreover, the renewed Danish focus on the Arctic is motivated by both NATO and the US, who have steadily pushed Denmark to improve the surveillance and defence of its own territory, as there is a growing concern over the so-called GIUK-gap that forms a naval choke point: the two stretches of ocean between Greenland, Iceland, and the United Kingdom. In 2021, this increased international pressure resulted in a political Agreement on Arctic Capacities (Danish Ministry of Defence 2021), that included acquisitions for DKK1.5bn (€201M) in surveillance and communication capabilities. This included air surveillance radars, coastal radars, space-based surveillance and satellite communication, drones, etc. The purpose of all these investments in new Arctic surveillance capabilities is to gather and process huge amounts of data necessary for intelligence, surveillance, and reconnaissance (ISR).

Spending on Intelligence Processing and Analysis only makes up a small part in the agreement, however, which indicates another challenge in the Danish military and strategic force generation culture: that is, hardware capacities are often prioritized above software and data processing solutions. This reveals a need for more efficient data management processes and software, allowing for better utilization of the gathered data through AI-enabled Image Intelligence (IMINT), e.g., for maritime surveillance, search and rescue, dark target detection etc. It is therefore important to discuss whether AI solutions should be developed by Danish industry and partners or whether Denmark should look elsewhere and use already developed solutions for data management.

Several players in the Danish defence industry are ready to contribute to developing such solutions and have explored these through workshops and exercises. One example is the ArcticX exercise (IRSA Development Group 2023) that was held in 2021 and repeated in 2022 at the H.C Andersen Airport outside of the Danish city Odense. The consortium IRSA (Integrated Remote Sensing in the Arctic) Development Group demonstrated how various sources of intelligence (based on sensor data from drones, satellite imagery, etc.) could be fused and integrated into a joint common operating picture (JCOP). Through use-cases and live demonstrations, the exercise showed how AI (Machine Learning and Deep Learning algorithms) could be used for dark target detection, fishing control, identification and tracking of icebergs, oil slippage, etc. However, there is also here a tendency to valorise platforms and capacities (such as sensors, satellites, and remotely piloted aircrafts) over software and the “invisible” systems that are supposed to integrate the gathered data.

The discussion of the Danish tendency to prioritize traditional hardware becomes more pertinent with recent procurements of advanced and costly capabilities for the Danish Airforce—such as several Seahawk helicopters and F-35 fighter jets. This is why the forthcoming strategic initiatives for a future digitalized Danish Defence allegedly aim to ensure coherence between operational and supporting areas, including the management of data. This is crucial if Denmark is to get to optimal output from new digital technologies and improve situational awareness on the battlefield as well as increase the speed of decisions, information sharing and communications. In short, the Danish Defence’s operational and administrative data as well as digital networks are critical to defence operations.

2 Developing Defence AI

In the pursuit of digital transformation, Danish defence has engaged in several research and development projects and activities related to AI. These involve partnerships with research institutions, collaboration with industry partners, focusing on developing AI algorithms, data analysis techniques, autonomous systems, and other AI-enabled technologies that can enhance military capabilities.

However, gaining insight into new military technology through research and development in the organisation has long been downgraded by broader political agreements and alternative military focus areas. In contrasts to both Norway and Sweden’s extensive military R&D capabilities, the Danish Defence R&D serves primarily to support the acquisition of new capabilities.

Consequently, the internal R&D of Danish Defence has been cut several times over the past 20 years while having been relocated from the services and centralized into the Defence Command under the Plan and Capabilities Division. Currently, Denmark spends less than 0.5% of its defence budget on research and development, compared to other EU countries at 1.7% (IDA 2023).

Independent R&D is therefore very limited and Danish Defence Research serves primarily as a node in a broader research network. This helps Denmark access allied cooperation partners and create new ventures with the defence industry (Breitenbauch and Mathiesen 2021).

2.1 Defence AI in Public-Private Partnerships

In the latest defence agreement for the period 2018–2023, research and development received renewed political attention. The agreement states: “The defence cooperation with industry on R&D is strengthened, among other things with a view to maintaining the Danish defence industry as an attractive partner for other countries’ defence industries” (Danish Government 2018: 11).

This trend is further strengthened with the political agreement for the coming period (2024–2033). The Government and a broad majority of the Danish Parliament have agreed that Denmark shall reach 2% of the Gross Domestic Product (GDP) on defence no later than 2030. The agreement entails investments in Danish security and defence amounting to approx. DKK143bn (more than €19bn) during the period 2024–2033. The government also plans to strengthen research and development in new defence and security technologies, including drone and quantum technology (Danish Ministry of Defence 2023: 10).

In 2021 the Danish Government released its National Defence Industrial Strategy as a capstone document on the subject. A central element in the strategy is the realization that “the Danish defence industry is essential for Denmark’s national security and our joint efforts with allies and partners” (Danish Government 2021: 5).

In short, Denmark finds itself in a changed and more complex threat landscape in which an increased great power competition is particularly pronounced in a long-term race for the development of new civil and military technology—such as the AI arms race between USA and China (Breitenbauch and Mathiesen 2021: 20).

The Danish AI R&D takes place in the private sector where technologies with commercial aims are merged into the military world. The strategy therefore articulates the need for stronger collaboration between the public sector, Danish research institutes, and the defence industry. It also calls for better utilization of the potential that exists in small and medium-sized companies and in innovation and start-ups sprouting from the research environment at universities.

With the intent of strengthening the cooperation for Danish security, the strategy invites both industry and academia to collaborate closely regarding defence AI research and development through new civil-military partnerships and a bolstered AI ecosystem.

2.2 The Danish AI Ecosystem

The Danish AI ecosystem consists of various organisations, initiatives, and stakeholders involved in the development, promotion, and application of defence AI. Denmark has been actively fostering its AI ecosystem to drive innovation, economic growth, and societal progress, which has been further strengthened by the National Defence Industrial Strategy.

The strategy specifically mentions AI as an area of focus: “The Government prioritizes providing scope for collaboration with companies regarding disruptive technologies, including artificial intelligence” (Danish Government 2021: 9).

Central to the AI ecosystem and the Government’s strategy are public-private partnerships and civil-military enterprises. This collaboration is based on the triple helix innovation theory, which emphasizes the importance of interaction between the public sector, industry, and universities. For instance, the strategy states how “innovation and new partnerships are necessary to ensure that defence authorities and Denmark’s defence industry are ready to meet the future” (Danish Government 2021: 20). It is crucial for the realization of the strategy that Defence is brought closer to collaboration partners from industry and Danish research institutes.

2.2.1 Defence

In the Danish defence organisation, research and development of defence AI is formed roughly through the so-called “Bermuda Triangle” between the Danish Ministry of Defence (MoD), the Defence Acquisition and Logistics Organisation (DALO) and the Defence Command. The annual briefing for the Defence Command describes, for instance, how the Chief of Defence Development Forum is used as an arena for discussions of new trends and opportunities across types of weapons and defence. The declared goal is to strengthen the Armed Forces by challenging conventional solution models by applying new insights and flows (such as IT as a “force multiplier,” including AI).

In addition, DALO is responsible for procurement, supply, maintenance, development and decommission of material capabilities, IT and services for the Danish armed forces and Emergency Agency. With the slogan “Open for Business”, DALO (and the Danish MoD, which launched it as a strategy in 2013) wish to signal openness and to support Danish industry and export, not least in the growing field of AI research and development (Danish Ministry of Defence Acquisition and Logistics Organisation 2012).

The Danish Defence Research Centre (Værnsfælles Videnscenter) supports DALO in research and development, trend analysis and technology scouting for current and future investments in military acquisitions, in particular on AI technology and other emerging and disruptive technologies (EDT). Moreover, the Danish Defence Research Centre co-finances research and development projects carried out in collaboration with Danish companies and research institutions.

In other words, the Danish defence uses co-financing projects to expand research capacity in several areas where it does not have the necessary resources at hand (due to the cuts on research and development). The purpose of the projects is partly to support the defence industry and partly to provide DALO with important technological knowledge for the benefit of the Armed Forces’ operations. These co-financing opportunities have become increasingly popular, especially after a stronger connection to the European Defence Fund (EDF) had been established.

A core element in the National Defence Industrial Strategy is to accelerate “speed, flexibility, transparency and enhanced triple-helix collaboration” as important parameters for this modernized acquisition practice (Danish Government 2021: 13). The strengthened civil-military cooperation and partnerships can be regarded as a significant paradigm shift in Danish military strategic culture. While several of Denmark’s allies, e.g., the Netherlands, France, and Norway, traditionally feature close ties between defence authorities and defence industries, the Danish Ministry of Defence and its agencies have historically been more reluctant to enter collaborations with the Danish defence industry. One reason is the historic Danish “defencist” way of thinking about military power in which private defence industry actors traditionally have not been invited into the governmental decision circles. It is therefore largely uncharted waters when the Armed Forces now enter various partnerships about AI with industry and research institutes.

2.2.2 Academia

Danish universities and research institutions play a crucial role in advancing AI knowledge and expertise. Institutions like the Technical University of Denmark (DTU), Aalborg University, University of Southern Denmark and the University of Copenhagen have research groups and programs dedicated to AI and machine learning. These institutions collaborate with industry partners and contribute to cutting-edge research. In particular, areas such as AI, big data, and quantum technology are targeted. Simultaneously, Danish research institutions are world-leading in select areas of space and military technology. This research can support Denmark’s national security by converting new knowledge into innovation in Danish companies and into solutions for Denmark’s operational entities.

For instance, DTU has extensive research in applications of AI for security purposes. And in 2022, NATO located a quantum research centre at the University of Copenhagen’s Niels Bohr Institute as part of the Defence Innovation Accelerator for the North Atlantic (DIANA) (The security policy analysis group 2022: 18). Another quantum technology test centre is located at DTU for the development and manufacture of quantum technological solutions. The centre develops quantum sensors and ultra-fast quantum encryption devices that can prevent hacking.

In 2023, the University of Copenhagen collaborated with the IT University of Copenhagen and other Danish universities to establish a Pioneer Centre for AI (P1). Also in 2023, the eight Danish universities and five government-approved technical service institutions (GTS) formed a joint National Defence Technology Centre (Nationalt Forsvarsteknologisk Center, NFC) to increase collaboration with industry and the Danish Armed Forces. Among the approved technical service institutions participating in the National Defence Technology Centre is the Alexandra Institute whose AI and Analytics Lab and Visual Computing Lab is one of the biggest commercially available providers of AI research and development in Denmark. The purpose of the National Defence Technology Centre is to contribute to a critical technological boost of the Danish Armed Forces and the Danish defence industry through interdisciplinary partnerships. It utilizes Denmark’s strengths in AI and similar areas such as space, quantum, cyber security, green fuels, autonomous systems etc.

2.2.3 Industry

In recent years, Danish private commercial entities—including companies such as TERMA, Systematic and Weibel—have actively explored the potential of AI in military applications to address various defence challenges and build partnerships with the Danish armed forces and academia. These companies focus on leveraging AI to enhance the capabilities of armed forces, improve efficiency, and contribute to national security.

The Danish private industry’s engagement in AI for the military is overall enthusiastic and encompasses a wide range of areas. One notable aspect is the development of autonomous systems, including unmanned aerial vehicles (UAVs) and ground vehicles. These autonomous platforms equipped with AI offer advanced reconnaissance, surveillance, and logistics capabilities.

Moreover, Danish companies specializing in defence AI have been at the forefront of developing intelligent data analysis tools and predictive analytics models for battle-management and command and control purposes. These technologies aid in processing vast amounts of data, extracting actionable insights, and supporting informed decision-making in real-time operational scenarios.

A significant, although paradoxical, example is the development of the F-35 fighter jet. For years, Danish defence companies such as TERMA have played a key role in the development and production of highly specialized AI-based software and sensor solutions for this fifth-generation fighter aircraft (Danish Government 2021: 13). Yet, the lack of an overarching Danish IT and data infrastructure has made critics doubt how the Danish defence organisation is geared to make proper use of the gathered data and sophisticated technologies built into the F-35 jets.

Another crucial aspect of Denmark’s private industry’s involvement in AI for the military is cyber security. As the digitization and interconnectivity of military systems increase, the need for robust cyber defence measures becomes paramount. Here Danish companies specializing in AI develop advanced cybersecurity solutions that employ machine learning algorithms to detect and mitigate cyber threats effectively.

2.3 Current R&D Projects on Defence AI

In 2023, TERMA, the country’s biggest defence company, set new standards for military R&D in Denmark by signing a 30-year contract with Danish Defence for delivering an air defence system to the Danish Army’s 1st Brigade and thereby meeting NATO’s capability goals for Denmark in Very Short-Range Air Defence (VSHORAD) (TERMA 2023).

With the agreement, TERMA becomes the responsible systems integrator for Danish air defence. This means that the company must integrate all the systems’ sub-elements and ensure that future air defence capabilities can form an integral part of the brigade’s operations in the overall air defence. This includes being able to communicate and exchange data with other air defence systems. The system will be based on TERMA’s AI-assisted command and control system (C2), which can connect several units via integration with various sensors and communication systems (TERMA 2023).

Another significant example of a current AI project and international partnership is the AI for Defence (AI4DEF). AI4DEF is a consortium consisting of 22 partners (companies and research institutions) from 10 countries, including Leonardo, Airbus, and the Danish Aalborg University and TERMA as the consortium lead. The AI4DEF’s intent is to develop defence AI in relation to specific use-cases, developed in collaboration with, among others, Danish Defence. According to the consortium, the aim is therefore to “pave the way for accelerated development and application of AI in defence in order to maintain European sovereignty and excellence in this area”. The consortium is backed and funded by the European Commission as a part of the European Defence Industrial Development Programme (EDIDP).

Situational awareness and decision-making as well as planning optimization are studied, designed and tested through the AI4DEF partnership. For instance, AI4DEF demonstrates how AI can help human operators exploit the increasing amounts of data/intelligence from various sources, including creating better situational awareness and situational understanding (e.g. supporting operational planning through modelling and simulation). This has been tested in one of several demonstrations, showing how AI can improve unmanned aerial vehicles for ISR missions through automatization and optimization of route planning and processing of unequalised data from a range of sensors and sources, e.g. satellite imagery, weather data, open-source intelligence, etc.

Several other R&D projects on defence AI have been co-financed by DALO. One example is a project that has prototyped and tested AI-based technologies for automated terrain analysis. Also, the Defence Command has run an experimental project on defence AI in collaboration with DALO. The project focused on testing and development of software using AI/Machine Learning to classify aircrafts based on radar data (Danish Defence Command 2022: 8–9).

3 Organising Defence AI

In Denmark, the lack of an overall organisational approach to defence AI is countered by upcoming strategic initiatives and a joint dialogue among different branches of the Danish Armed Forces, the Ministry of Defence, the Cyber Division (part of DALO), the Defence Command, and other relevant stakeholders. The primary objective is to leverage AI technologies to enhance the efficiency, effectiveness, and safety of defence operations. Only recently, Denmark has started to reorganize and adjust the organisational structure for a better adoption of defence AI. The formation of the above-mentioned Cyber Division in 2023 remains central to the development of the organisation’s overall IT architecture and communication infrastructure.

At the organisational level, it is crucial for the overall readiness and maturity of the organisation to develop a so-called “digital backbone” for the defence collaboration and information sharing among the different services and stakeholders. While it has not yet been specified what this “digital backbone” should include, it has become clear that it is not primarily a technical solution or system but rather a concept, including not only technology or data, but also people and processes.

3.1 Joint Approaches

At the ministerial level, a forthcoming capstone document on digitalization will allegedly be decisive in setting a direction for the digital transformation of the organisation. The actual organisational changes that result from the digitalization strategy are expected to take place gradually and iteratively during the period from 2024 to 2029. In this regard, it is important to develop an open and agile mindset that allows for testing of AI-driven initiatives (or within other EDTs), which can be shut down (“fail fast”) if they do not provide the desired value. Moreover, the forthcoming strategic initiatives and capstone documents will supposedly have a strong focus on data exploitation including the future users of digital technologies and AI systems as well as the design of user-friendly and intuitive AI-driven solutions. In short, the users must be involved through user development and implementation and by using agile methods.

Preparing the organisation, including its people and culture, for the successful adaptation of defence AI, the Danish Ministry of Defence also participated in international fora and working groups, including the British-American initiative AI Partnership for Defence (AIPfD). Besides partaking in seminars and meetings in the partnership, Denmark also contributed to a sub-working group with the aim of developing an AI Readiness Model. In particular, Denmark contributed to the readiness framework with one out of the eight building blocks, which can be used to measure the AI readiness in the defence organisations: namely that of Culture and Organisation.

3.2 Single Service Approaches

The Danish Armed Forces aim to integrate AI into various aspects of their operations. The individual services within the Danish Armed Forces, such as the Army, Navy, and Air Force, have their own specific initiatives and applications for AI.

While the Army focuses on using AI for autonomous systems such as drones and other unmanned aerial or ground vehicles (UXVs) or intelligent logistics systems, the Navy explores AI-driven systems for maritime surveillance and autonomous underwater vehicles (DALO 2021b). Also, the air defence systems integrated on the Danish frigates such as CIWS (close-in-weapons-systems) and the newly procured SM2 air defence missiles are examples of AI-based systems incorporated into military technologies (Naval News 2021).

The Air Force leverages AI for tasks such as aircraft maintenance (e.g., predictive maintenance), mission planning, air traffic management and command and control—all of which are necessary for organising and operating the newly procured F-35 fighter jets.

In addition, the Danish Emergency Management Agency (DEMA) under the Ministry of Defence is already using AI-enabled systems for pattern analysis based on drone imagery in emergency operations and for search and rescue (SAR). Moreover, DEMA has strengthened its digital support and integration with international and national partners, such as the municipal emergency services and the police.

3.3 Building an AI-Driven, Data-Centric Organisation

For several years, the Danish defence has worked on a common data and information infrastructure, now called “the Danish Common Operational Information Environment” (DACOIE). The concept is supposed to provide the basis for an operative and data-based overview that strengthens the ability to make faster and better decisions and communicate them across defence organisational units.

While there has been some initial dispute concerning whether DACOIE should be regarded as a technical system or more of an idea or concept, the latter approach now seems to dominate. Thus, DACOIE represents a general direction for how to build a digital backbone for the organisation that enables the Danish armed forces to operate digitally and integrated across all five domains (land, sea, air, cyber and space), and allows live data to be exchanged between domains and units.

The incentive behind DACOIE is also based on needs and demands from Denmark’s allies and partners, most importantly NATO, the US, and the UK. For a while, the conceptual discussions in the Alliance have focused on digital transformation to secure its position in the future. This includes converting its data usage to cloud-based solutions, where AI-enabled data analysis remains central as a basis for decision-making that outperforms that of our adversaries. The construction of a digital platform (Defence Information Cloud, DEFIC) is thus a prerequisite for this collaboration. And this is probably why DACOIE has become a prestige project for Danish defence—which gives the words from the Danish Chief of Defence about “servers before tanks” another dimension.

4 Funding Defence AI

With the Danish defence agreement for 2024–2033, the Government and a broad majority of the Danish Parliament agreed that Denmark shall reach 2% of GDP for defence no later than 2030. The agreement entails increased investments in Danish security and defence amounting to approx. DKK143bn (more than €19bn) during the period 2024–2033.

The government also plans to strengthen research and development in new defence and security technologies, including not only drone and quantum technology but also development of defence AI. As previously mentioned, this is a continuation of the latest defence agreement for the period 2018–2023, in which research and development received renewed political attention. As a rough assessment of the spending on digitalization and defence AI, the defence budget for 2018–2023 contained a range of new initiatives, including spendings on cyber and IT security for DKK567M (€76M).

The Danish government supports research and innovation projects related to AI and defence technologies, e.g. Innovation Fund Denmark and The Danish Council for Research and Innovation Policy. Moreover, Denmark has a growing technology sector and an active venture capital ecosystem that supports various industries, including defence and AI. The Danish Government has launched a range of initiatives to strengthen the national AI ecosystem, which could also benefit the development of defence AI. Prominent venture capital firms and investors in Denmark that have shown interest in AI and defence-related sectors include Denmark’s Export and Investment Fund (EIFO) (previously Vækstfonden, AI Denmark, and Nordic Eye Venture Capital).

5 Fielding and Operating Defence AI

Defence AI will be central to future Danish operations as part of an information environment where the right information is available to the right people, in the right format, at the right time. The newly established DACOIE concept aims to improve the organisation’s communication infrastructure and ability to process data and information to use AI to achieve battlespace advantage and connect sensors to decision-makers and effectors (supporting the understand-decide-act functions).

Danish defence is already working with AI-enabled command and control (C2) solutions. This includes the SitaWare Suite developed by the Danish company Systematic, which is used by the Danish Army as the preferred battle management system to create situational awareness and to support military leaders in decision-making on the battlefield. While the use of AI in SitaWare has so far been limited, the latest add-on to the suite, SitaWare Insight, further increases the use of AI for detection of anomalies with algorithms trained for detecting and assessing deviations from the common operating picture. However, the add-on has not yet been purchased by the Danish Defence, which means that the armed forces are still missing the AI-driven software to make proper use of the increasing amounts of data and information made available on the battlefield.

As discussed above, the discourses, narratives and imaginaries surrounding defence AI in Denmark—whether it is ethical principles in the National AI strategy or simply a lack of AI literacy and digital skills among the people in the organisation—can make it hard to imagine how AI is used in Danish defence. Yet, automation and AI-assisted solutions are already enrolled to some degree in the organisation for several different purposes, as discussed below.

5.1 Administration and Logistics

Administrative robots, or RPAs (Robotic Process Automation), have been used by the Danish defence since 2019 where they were created by the Robotics Operational Centre, part of DALO. The software was developed to automate administrative tasks, allowing military employees to focus on core tasks. Feedback from employees has indicated that using AI for the automation of workflow can make a big difference. Not least on one of the busiest administration days of the year in the Danish defence, namely the conscription day for the Army. Here, the robots handle large amounts of data and ensure that the conscripts are quickly registered in the Defence systems. The enrolment of administrative RPA robots for handling conscription data is expected to inspire several other administration and logistics functions in the organisation (DALO 2021a).

5.2 C4ISR

Various AI-based types of decision support in Command, Control, Communications, Computers, Intelligence, Surveillance, and Reconnaissance (C4ISR) systems are currently used by Danish defence. Systems such as the SitaWare suite by Systematic and the JIMAPS from TERMA incorporate various AI techniques and technologies to enhance the situational awareness of military decision-makers. For instance, AI is used to integrate and fuse data from multiple sources such as sensors, surveillance systems, satellites, and intelligence databases. The use of AI for such purposes will assumable be more broadly implemented if or when system add-ons (like SitaWare Insight) or a common data and information infrastructure is rolled out. This will also enable other AI techniques like data mining, pattern recognition, and machine learning to identify relevant information, filter out noise, and provide a comprehensive view of the operational environment.

5.3 Unmanned and Autonomous Systems

In recent years, the Danish military has started testing and using unmanned systems in a tactical capacity for tasks like reconnaissance and surveillance, target recognition, route planning, autonomous navigation, and mission execution. AI is increasingly incorporated into these unmanned or autonomous systems that enable effective coordination and collaboration between human operators and autonomous systems, enhancing overall ISR capabilities. For example, the Danish Army plan to integrate surveillance and reconnaissance drones into their units on all tactical levels expected to rely on AI in mapping terrain or threats and in processing data through on-board computing.

In the maritime domain, AI is about to be fielded on underwater drones and sonars to monitor and protect critical infrastructure on the seabed. Systems such as the Katfish highspeed towed sonar produced by Kraken Robotics are examples of how AI or Synthetic Aperture Sonar (SAS) enabled systems can help monitoring the seabed. Also, six new underwater drones (Light Autonomous Underwater Vehicle (LAUV)) from the Portuguese company OCEANSCAN have been procured and is to be used by the Danish Naval Command’s Mine Counter Measure unit (MCM) on Mine Sweeping Drone ships (MSD). The new LAUVs are used when the MCM unit must search for and detonate mines on the seabed. In their search for mines, the units can be deployed simultaneously and use AI to calculate how to search the seabed within a given area as well as communicate with each other so that if one finds something, it autonomously calls another unit to take a photo of the object.

5.4 Cyber Security

As mentioned, Denmark is one of the most digitized societies in the world. Many critical sectors in Denmark are digitized and essential for ensuring the functioning of society. This explains why cyberattacks are particularly threatening to Denmark. In the National AI Strategy, the Danish Government states that it is a priority that Denmark preserves and develops national cyber expertise capable of protecting Danish society and the Danish defence industry against external attacks. This is done by using intelligent cyber security solutions such as the QRadar program developed by IBM. Using AI, QRadar can examine billions of pieces of information in a short period of time and look for signs that a network or system has been compromised by malicious players. The program can therefore help IT specialists find breaches of security that would be hard for a human to detect (Danish Government 2021: 21).

6 Training for Defence AI

Another way to “fill the gap” would be to prioritize and increase the teaching and training in AI and technology management in both current and future military education such as at the military academies and in the Master’s in Military Studies program offered by the Royal Danish Defence College.

Currently, Danish Defence is not exploring all the possibilities that AI can offer for military education and training. As an example, the Command Course at the Royal Danish Defence College uses simulated exercises as an element in the training of the Danish Army’s future staff officers at the battalion and brigade level. Using different types of simulation tools during the command exercises, the students get acquainted with the different roles and routines of staff work while both planning and decision-making are practiced.

In the digital part of the exercise, all input is generated through the more than 20-year-old simulation program Steel Beasts. The digital simulation in Steel Beasts is created based on the students’ own staff work and results in a plan that is supplemented during the simulation. Through this combination of analogue and digital input, the students are exposed to a simulation that is unknown to them, and where the possibility of unforeseen actions on the part of the opponent is greatest. Yet, the element of surprise and unexpected Red Team manoeuvres are relatively limited due to the outdated AI software in Steel Beasts. But if combined with a more up-to-date AI computer simulation that cannot be controlled in the same way as the input during the existing computer game, then the mutual actions and decisions made by the students could have a much greater influence on the situation at all levels.

The Steel Beasts example serves as a symptom of how technologically immature the Danish military education system is in incorporating cutting-edge AI technology into existing courses, training, and exercises. However, some early-stage initiatives are starting to sprout at the Royal Danish Defence College: there are ideas of creating an AI lab for experimenting and introducing military officers to the possibilities enabled by AI and Machine Learning. Such initiatives should serve as prototypes for teaching programs in AI technology at all levels of the military education system.

First, Danish defence personnel should receive specialized training and education in AI concepts, principles, and applications. This includes understanding the fundamentals of machine learning, data analysis, algorithm development, and AI ethics. Training programs may involve courses, workshops, and certifications to equip military personnel with the knowledge and skills required to work with AI technologies.

Secondly, Danish defence should establish stronger partnerships with academia and industry, including dedicated AI centres or labs where experts and users can conduct research, experimentation, and development of AI. These centres could serve as hubs for innovation, fostering collaboration, and providing resources for training personnel in AI. And they should facilitate hands-on learning, prototyping, and testing of AI algorithms and user-friendly systems specific to defence applications.

This should also involve simulation and training exercises, which are crucial for familiarizing defence personnel with AI-enabled systems and their operational implications. These exercises should create realistic scenarios where personnel can practice using AI tools, analyse AI-generated data, and make decisions based on AI recommendations. Training exercises also help identify any limitations, challenges, or vulnerabilities associated with AI implementation in defence contexts. But most importantly, training is necessary to build trust in AI systems and make military personnel feel comfortable with the transition from a human-centric organisation into human-machine teaming and “centaur”-warfighting (Warren and Hillas 2020).

Finally, the future training for defence AI in Denmark should emphasize ethical and legal considerations surrounding AI usage. This includes understanding the potential biases (including gender and race), risks, and consequences associated with AI, ensuring compliance with national and international laws, and adhering to ethical principles for responsible AI development and deployment. Training programs should address the ethical use of AI in military operations, data privacy, and the protection of civilian rights.

7 Conclusion

Denmark is still at an early stage of adopting defence AI. Despite the government’s vision for Denmark to be an AI frontrunner. The National Strategy for Artificial Intelligence does not provide specific guidance for the military’s implementation and use of AI, focusing more on public-sector services and growth in the business community.

Despite this, AI is being tested and used in the Danish defence for some purposes, including “low-hanging fruits” such as administrative tasks and logistics as well as in training and education. The problem is, however, that the exploration and initial testing of AI systems is highly fragmented and isolated within silos of the organisation. This results in redundancy and problems with sustaining the systems, challenging a near-future integration of AI into C4ISR systems that can enhance decision making and situational awareness.

Danish defence collaborates with the defence industry and academia regarding new partnerships and research of defence AI. Unmanned and autonomous systems with AI capabilities are being tested and evaluated for tasks such as intelligence, reconnaissance, and surveillance operations where AI is used for automation purposes such as creating waypoints and calculating routes.

However, there is a growing knowledge gap and lack of AI literacy in the Danish Defence, which makes the recruitment of IT specialists and data science experts an increasingly demanding task. Although the Danish Chief of Defence has emphasized the importance of investing in AI and data processing software, the real challenge will be to recruit or educate the right people with the right skills and competences to make proper use of these AI-enabled data and software systems. And in this endeavour, it remains pertinent to create a culture of thinking about technology, including defence AI and data, not as barriers but as something that the organisation can profit from.

The need for efficient data management and software systems has become evident with the procurement of advanced military capabilities. Danish Defence recognizes the criticality of data and digital networks for operational success, and future operations are integrated across multiple domains. Yet, the increasing amount of data gathered from advanced sensor platforms is not always used properly because it rarely reaches further than the crew in a fighter plane or frigate. Therefore, Danish Defence needs to adopt AI as a part of the development of a much stronger digital backbone to process the increasing amount of information and share it across platforms, units, and domains. Especially the latter is important if Denmark is to remain relevant for its allies in NATO and abroad.

While Denmark acknowledges the importance of defence AI and digitalization, there are huge challenges to overcome. Above all, the growing knowledge gap and lack of digital literacy is considered a major risk and should be met with investments in education and training. The Danish Defence aims to leverage AI to improve operational capabilities and decision-making processes, but careful consideration is required to ensure responsible and explainable use of AI in alignment with international norms and standards.