The newest REMIT dashboard is based on the research of our Finnish partners, Finnish Institute of International Affairs (FIIA), and explains how military AI is handled globally in the lights of two key frameworks: Responsible AI in the Military Domain (REAIM), and Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy.
Here is how the FIIA team introduces their dashboard:
“At present, there is no shared framework addressing the evolution, deployment and use of Artificial Intelligence (AI) in the military domain. Two incipient initiatives emerged in 2023: the Responsible AI in the Military Domain (REAIM) Call to/Blueprint for Action and the Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy spearheaded by the United States. Although the two initiatives emerged concurrently at the first REAIM Summit, the processes have diverged since and the groups of countries that have signed onto them are not congruent.
While the REAIM Call/Blueprint opts for a multi-stakeholder approach to tackling the challenges of AI in the military domain, the Political Declaration focuses on states and can be viewed as a coalition- and confidence-building instrument.”
Responsible AI in the Military Domain (REAIM)
The first summit on Responsible AI in the Military Domain (REAIM) took place in the Netherlands at the World Forum in the Hague on 15 and 16 February 2023, gathering stakeholders from governance, civil society, industry, universities and think tanks. Experts assembled with an objective to build a common understanding on opportunities and risks in military AI.
While AI was outlined as creating “unprecedented opportunities” particularly in its assistance to decision-making, its development was also seen as raising legal, security and ethical concerns for social transformation, concerns that are magnified in the military context.
REAIM 2023 aimed to: politically elevate responsible approaches to military AI; energize a broad stakeholder community; and share experiences, best practices and solutions.
The summit centred on three core themes: mythbusting AI – to better understand AI’s technicalities and explore how these can be responsibly utilized in military environments; responsible deployment and use of AI – exploring the benefits and vulnerabilities of military AI applications in practice; and governance frameworks – to identify existing AI governance frameworks and explore additional opportunities for stronger frameworks.
REAIM 2023 concluded with a Call for Action supported by over 50 countries. This notably included the main Western powers and China, but not Russia. It affirmed that the most important developments shaping AI happen in the civilian domain while calling for a multi-stakeholder outlook on challenges posed by military AI.
A second REAIM summit was held in Seoul, South Korea in September 2024, finalizing a subsequent Blueprint for Action supported by over 60 countries on the functions of AI in global security; responsible AI in military contexts; and emerging governance frameworks for military AI.
The Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy
The Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy (the political declaration) was launched by the United States during the Biden presidency, at the Responsible AI in the Military Domain Summit (REAIM) in the Hague on 16 February 2023.
The political declaration calls, inter alia, for responsible development, deployment, and use of AI capabilities, legal reviews to ensure compliance with international legal obligations, proper oversight, minimization of unwanted bias, and taking appropriate care in introducing AI to military systems. It also underlines the development of understanding regarding the functioning of military AI capabilities along with implementing appropriate testing, assurances and safeguards. The first plenary meeting for the Political Declaration was hosted by the US on 19 and 20 March 2024 to converse on national approaches, priorities and challenges. The meeting also set up three working groups with the aim of producing specific deliverables down the line.
The political declaration contains a developing non-binding normative framework for the use of AI capabilities in the military domain. It is a confidence-building measure, which allows state parties to share best practices and voice their support through endorsement. The novelty of the declaration is its remit, which includes all military uses of AI – not only lethal autonomous weapons systems (LAWS). While principally open to all states and explicit in calling for endorsing parties to support other initiatives in the military AI governance space, the declaration has been criticized for being a vehicle for the US to pursue its strategic interests with likeminded allies.
It is unclear at present whether the objectives of the political declaration will be furthered by second Trump administration, given its profound aversion to multilateral frameworks, global governance initiatives and the regulation of AI in general.
References for the data:
U.S. Department of State (May 29, 2024). “Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy”.
https://www.state.gov/political-declaration-on-responsible-military-use-of-artificial-intelligence-and-autonomy/.
Government of the Netherlands (Feb 6, 2023). “REAIM 2023 Endorsing Countries and Territories”. https://www.government.nl/documents/publications/2023/02/16/reaim-2023-endorsing-countries.