The rapid military integration of artificial intelligence, occurring against a backdrop of geopolitical fragmentation, is increasing the risk and unpredictability of “flash wars.” This regulatory vacuum is further compounded by tensions between, on the one hand, state power and, on the other, the technical control capabilities of private actors.

Indeed, for example, the fact that the Pentagon designated Anthropic as a “supply chain risk”—in an act of obvious retaliation against the U.S. company’s ethical safeguards—is a clear testament to the asymmetries in regulation and control over technology. Furthermore, regarding the military use of Artificial Intelligence, the reliance on probabilistic and inherently fragile models (LLMs), integrated into decision support systems (AI-DSS) without rigorous reviews under Article 36 of Additional Protocol I to the Geneva Conventions, introduces systemic vulnerabilities.

If the legality of the use of force becomes elastic and unilaterally defined by the executive branch in confrontation with the private sector, how can we ensure distinction and proportionality in the context of armed conflict? Are we institutionalizing algorithmic impunity under the pretext of national security?

Speakers: Azeredo Lopes – Faculty Member, FD UCP Porto, and Pedro Braga de Carvalho – Visiting Faculty Member and PhD Candidate, FD UCP Porto