Intelligent armies: how can AI influence the CSDP?

During this year, hugely affected by the pandemic, Artificial Intelligence (AI) has become one of the biggest players. The defence sector itself, which is often not considered as a European issue, is having significant changes, both in terms of arms and operations. The United States and China have already made significant investments in AI. By contrast, the European Union (EU) is still looking for direction. What if the EU exploit this moment to become a real technological sovereign power?


AI will have a major impact on our society for years to come. Its applications in the defence sector will be various as well, and it is not hard to see why AI could represent a great chance as well as a great danger at the same time. It will be used in target recognition, in the cyber warfare domain, in logistic and transport, in wargaming, in data information processing, and in the development of new weapons systems. The automation and use of such technology would also lead to faster progress in the military decision-making process. Nevertheless, the lack of human oversight could affect and ethics both in arms and operations[1]. It is important not to consider AI as something far in the future, but a present issue which should be taken into account and, if necessary, controlled.
AI might represent the real opportunity for the European Union (EU) to boost its Common Security and Defence Policy (CSDP), especially in risk detection, in the protection and preparation capabilities, and the improvement of the EU’s defence production capacities. As also expressed by Jean-Francois Ripoche, Director for Research, Technology and Innovation of the European Defence Agency (EDA), AI may represent a source of opportunities and vulnerabilities at the same time.

Regarding the opportunities, Riproche highlights that AI would bring to a risk reduction of losing life in conflicts, to a growth in the efficiency of equipment and people, to a decrease in the costs of training and operations, to an improvement in intelligence gathering and better protection for soldiers and civilians during conflicts. On the other hand, he also expresses how many threats and dangers AI may cause, such as manipulation of information, Cyber-attacks, and trust-related problems. It is essential, indeed, to find an equilibrium in the relation AI-defence, and to develop an approach based on trust and confidence between AI technologies and soldiers[2]

But how AI could be concretely used in CSDP operations? AI could have three main uses: detection, preparation, and protection. New technological systems could indeed help the EU’s foreign policy to understand and gather relevant data, to boost EU’s capacities at the tactical level, favouring peacebuilding and humanitarian missions (especially in post-conflict or fragile areas), to tackle and identify to disinformation and defend the cyber-domain, by revealing fake images, videos and audio files, as has been partly shown during the COVID – 19 crisis[3],  and finally to help improving border’s assistance initiatives through biometrics or facial recognition, which may be used alto to counter fight terrorism[4].  Nevertheless, the use of data by facial recognition systems is still debated in the Brussels Bubble, as the Clearview AI case and concerns raised by the civil society have shown[5].

One practical example of the potential of AI in the military environment is given by the case of the EU Satellite Center (SatCen). During this last year, SatCen has developed an automatic process, powered by AI, which collects a large amount of data through satellite, convert them and make them available for further analysis. This is only one of the various processes being tested in conjunction with the European External Action Service, the EU Military Staff and the EU Intelligence and Situation Centre[6]. Those processes may have a strong impact on CSDP operations, helping the military staff to have a better comprehension of the territory where the personnel are deployed. 

AI application in CSDP operations


Source: D. Fiott, G. Lindstrom, Artificial Intelligence: What implications for EU security and defence, European Union Institute for Security Studies, cit.

Particularly the EDA will manage AI in the European military environment, having a key role in the Research and Development (R&D) and the coordination among European armies: the concept of AI is very fluid, and it was interpreted differently by member states. To overcome these discrepancies, the main tool has been the Permanent Structured Cooperation (PESCO). Since 2018, the EDA launched several projects to exploit the potential of AI[7]. Also, the Agency has developed from February 2019 a strategy based on three phases. The first and the second phase, ended in 2019, aimed to set a common understanding of AI in the military environment and to identify and analyse the various applications in R&D[8]. The last phase, which will last until the end of 2020, aims to create a new AI action plan, based on the needs and requirements of European countries[9].  This approach could bring many advantages, face the challenges of data sharing and control over AI, and attract talent innovation.  The Agency is indeed working to become a hub for exchange among European countries and to create synergies with other EU institutions.

The difficult relation elapsing the EU, AI and defence, may affect technological military development. The EU should truly understand which role to have in the AI race, and how to define itself in the geopolitical framework. As a matter of fact, many divergences emerged among the two historical transatlantic allies, especially in the digital domain: AI, 5G networks, and Big Tech regulations, are just some issues that will not be solved in the next future[10].

Despite the focus of the Commission guided by Ursula von der Leyen, aiming to bring Europe to the next level on foreign policy, may represent a good sign, the completion of a full European technological sovereignty is still far. Surely the EU has taken first steps to improve its AI development in the public and private sector[11], but the same does not apply in the defence one. In addition, the COVID – 19 crisis hit hard the EU, leaving military ambitions behind: the budget for the European Defence Fund, initially foreseen of €13 billion in June 2018, was deeply revised downwards up to €7 billion. As result, technological investments in the military sector will be slower than foreseen.

As now the EU position in the new global order is at risk, the Union should exploit this opportunity and fill the technological gap and develop an advanced AI Strategy also in the military environment. The strategy and projects developed by the EDA clearly marked a new will to explore all the potential of technology in security and defence, but contradictions remain. Despite European governments and armies know that cooperation is essential to advance in their technological defence, they are not yet ready to invest more in their security.

Therefore, it is essential to continue the work defined by the EDA’s action plan and use all means provided by the CSDP. In particular, the Particularly the Permanent Structured Cooperation (PESCO) has and will have a clear role in such process. In this direction, the Strategic Compass, the last doctrine developed under the German Presidency[12], should represent a new momentum for a closer cooperation among member states and for the birth of a European digital defence. Having the German locomotive in charge will indeed influence other member states, as shown during the debate over the Next Generation EU.

Eventually, rather than anything else, is to change the global imaginary on AI, starting to consider it as a trustworthy tool. Changing this picture will require strong efforts, and all the actors involved in the decision-making process should play their part.


[1] See D. Fiott, G. Lindstrom, Artificial Intelligence: What implications for EU security and defence, European Union Institute for Security Studies, November 2018, pp. 1-2.}

[2] See Presentation by Jean-François Ripoche at SEDE public hearing on Artificial Intelligence – enabled systems in security and defence,}

[3] See S. Tong, “Coronavirus: Can artificial intelligence be smart enough to detect fake news?”, Marketplace, 12 February 2020

[4] See N. Vinocur, “French politicians urge deployment of surveillance technology after series of attacks”, POLITICO Europe, 30 October 2020.} 

[5] See V. Manancourt, “U.S. facial recognition technology likely illegal in Europe”, POLITICO, 6 October 2020; J. Delcker, “Activists urge EU to ban live facial recognition in public spaces”, POLITICO Europe, 12 November 2020

[6] For a broad vision of SatCen activities involving AI, see European Union Satellite Centre Annual Report 2019,

[7] For further information on the current projects ongoing, see

[8] See European Defence Agency Annual Report 2019,

[9] See European Defence Agency, “Artificial Intelligence: Joint quest for future defence applications”, European Defence Matters, Issue 19 (2020), pp. 34-36,

[10] See N. Vinocur, “Europe and the US are drifting apart on tech. Joe Biden wouldn’t fix that.”, POLITICO Europe, 3 November 2020; M. Peel, H. Warrell and Guy Chazan, “US warns Europe against embracing China’s 5G technology”, Financial Times, 16 February 2020

[11] See European Commission, White Paper on Artificial Intelligence – A European approach to excellence and trust, COM(2020) 65 final, Brussels, 19 February 2020

[12] See Germany’s Presidency of the Council of the European Union, Strategic Compass: Developing strategic principles, 25 August 2020,

Foto copertina.

[vc_btn title=”Scarica Pdf” style=”classic” color=”primary” size=”lg” align=”center” css_animation=”bounceIn” link=”|||”]