The European Parliament's Subcommittee on Human Rights and Committee on Security and Defence on 15 July 2026 held a joint workshop on human control and AI-enabled military systems, revealing deep divisions over whether to prohibit or conditionally use lethal autonomous weapons, and whether existing international humanitarian law suffices or must be supplemented by human rights law. The session, opened by Mounir Satouri (Greens/EFA), also featured an exchange on a draft own-initiative report on AI and human rights in EU external action presented by Nacho Sánchez Amor (S&D).
Seven key disputes emerged. First, the legal framework: Netta Goussac (SIPRI) argued military AI must be assessed under human rights law, not only international humanitarian law (IHL). Second, the definition of control: Jessica Dorsey (Utrecht University) contended that nominal human presence is insufficient without time, authority, and information to exercise meaningful control. Third, the speed and scale of AI-enabled targeting: Katrina Manson (Bloomberg) reported that Project Maven accelerated kill chains from under 100 to 1,000 targets daily, raising concerns about legal judgment. Fourth, the reliability of AI: Manson cited accuracy drops from 70% to 30% in Ukraine, challenging the precision narrative. Fifth, the need for new standards: Goussac recommended parliamentary action and budget scrutiny. Sixth, total prohibition vs. conditional use: Lynn Boylan (The Left) aligned with the UN Secretary-General's call for prohibition, while Michael Gahler (EPP) defended contextual differentiation. Seventh, strategic necessity vs. rights-first guardrails: Christophe Gomart (EPP) framed defence AI as a strategic priority.
A European Commission representative clarified that the AI Act excludes military systems due to its internal-market legal basis, with defence governed by public international law. An EEAS representative outlined the EU's human-centric, human-rights-based AI approach in multilateral security work. Despite the divergences, consensus emerged on AI's current military use, the need for human control, upstream accountability, and Parliament's scrutiny role. Sánchez Amor indicated negotiations should begin after summer, aiming to advance the file by year's end.
The debate pits human rights and civil society advocates, who favour a broad human-rights lens and prohibition, against defence industry and military strategists, who prioritise strategic necessity and conditional use. EU institutions face pressure to reconcile internal-market AI rules with external action, while member states may resist new binding standards. The outcome will affect the EU's credibility in multilateral forums, where it advocates for human-centric AI, and its defence autonomy, where AI-enabled systems are seen as a strategic imperative.