Antonio Tejeda Encinas. President of the Euro-American Committee of Digital Law - CEA Digital Law
The EU Artificial Intelligence Act (AI Act), which entered into force on 1 August 2024 and will apply for the most part from 2 August 2026, is a crucial step forward in the regulation of emerging technologies, with a clear focus on protecting fundamental rights, particularly privacy and security. One of its most debated features, however, is the explicit exclusion of military and national security uses, which leaves a significant regulatory gap that deserves careful consideration.
The military exclusion: Article 2(3) of the AI Act
Article 2(3) of Regulation (EU) 2024/1689 provides that the Regulation does not apply to AI systems used exclusively for military or national security purposes. As a result, AI systems intended for defence are not subject to the strict restrictions imposed in other sectors. The exemption also extends to the output of AI systems used for military purposes that are not placed on the market or put into service in the European Union.
Recital 24 of the preamble reinforces this position, stating that the rules protecting fundamental rights in civilian uses of AI do not extend to military applications. This reflects a strategic decision by the EU not to restrict the development and deployment of AI in the military sector, which is regarded as essential to the security and strategic autonomy of its Member States. Yet the same decision raises concerns about the lack of oversight and the potential misuse of advanced technologies in defence settings, with possible unforeseen consequences for human rights and global stability.
The international perspective and regulatory efforts
Despite the exclusion in the European Regulation, the military use of AI has not been left entirely unregulated at international level. At supranational level, NATO has taken a proactive approach to regulating AI in military applications. In October 2021, NATO Defence Ministers approved an ‘Artificial Intelligence Strategy’ built on six core principles: lawfulness; responsibility and accountability; explainability and traceability; reliability; governability; and bias mitigation. These principles are intended to ensure that AI is developed and used in defence ethically and within a framework of clear accountability.
To ensure compliance with these principles, NATO has set up the ‘Data and Artificial Intelligence Review Board (DARB)’, a multidisciplinary body responsible for overseeing the use of AI in the military sphere. The Board is also tasked with developing certification standards so that the defence industry and the armed forces deploy AI systems that comply with those principles. This work is complemented by the European Parliament’s repeated insistence on human control over autonomous weapon systems, a concern reflected in several resolutions, including that of 12 September 2018.
Within the United Nations, moreover, the recent General Assembly resolution (UNGA 78/241) underlines the importance of global regulation of autonomous weapons, stressing the need for rigorous control to prevent the indiscriminate or irresponsible use of these technologies.
Regulation in Spain: a national example
Spain has made significant progress in regulating the military use of AI at national level with the approval of Resolution 11197/2023, which sets out the ‘Estrategia de Desarrollo, Implementación y Uso de la Inteligencia Artificial en el Ministerio de Defensa’ (Strategy for the Development, Implementation and Use of Artificial Intelligence in the Ministry of Defence).
The strategy adopts the six principles formulated by NATO and adds a seventh: ‘privacy’. This principle underlines the need to respect privacy at every stage of the development and use of AI in military settings, reflecting a comprehensive approach to the protection of fundamental rights.
The Spanish strategy also gives priority to the use of AI in several key areas, such as military mobility, intelligence, cyber defence and decision support.
To ensure effective follow-up, a ‘permanent working group’ has been created to develop and oversee the strategy. The group will advise the authorities on the ethical issues arising from the use of AI in the military sphere, and will also promote research and training in AI technologies for military personnel.
Ethical and legal considerations
Excluding the military use of AI from the European Regulation poses considerable ethical and legal challenges. Principle-based regulation, such as that adopted by NATO and Spain, offers a guiding framework, but it lacks the precision that more specific and detailed rules could provide. This ambiguity may lead to divergent interpretations and inconsistent application of the rules, which in turn could result in unintended risks or the misuse of AI technologies in the military sphere.
In short, while the European Union has made a significant effort to regulate civilian uses of AI, its decision to exclude military uses from the Regulation leaves a regulatory gap that raises ethical and security challenges.
This analysis is based on the text of the AI Act as currently drafted and published in the Official Journal of the European Union, and on the discussions surrounding its implementation.
Newsletter
Events, news and analysis on digital law, straight to your inbox. Only when we have something worth telling you.
