Military AI companies and their employees generally retain civilian protections under international humanitarian law, but that status is not absolute. In an analysis published by Eurasia Review, SIPRI researchers Laura Bruun and Marta Bo examine when people, assets and companies supporting military AI could face different legal consequences.
Watch Desk analysis
What happened
The analysis looks at three connected questions: when company personnel or assets might become lawful military targets, when workers or companies could face liability for violations of international humanitarian law, and what responsibilities states have towards suppliers.
It says employees could lose civilian protection if they directly participate in hostilities, while company assets could become military objectives if they meet the relevant legal test. Applying those tests to technical support, AI systems and dual-use data centres can be difficult, the authors argue. The analysis also notes that criminal liability depends on the applicable law and the individual or company’s involvement and knowledge, with significant legal and evidentiary hurdles.
The authors place primary responsibility for regulating suppliers on states, including through weapons reviews and clearer guidance on industry duties. Read the analysis.
Why it matters
Military AI depends on more than the system itself. Companies may provide ongoing maintenance, updates, training or advice, making it harder to draw a neat line between supplying a tool and supporting its use. The legal questions matter for workers and firms, but also for states deciding how to procure, oversee and separate military and civilian infrastructure.
Our read
The useful contribution here is a map of difficult questions, not a new verdict on any particular company or conflict. SIPRI’s analysis stresses that legal tests depend on the facts and that accountability can be hard to establish. That is a less tidy story than “AI firm liable” or “AI firm immune”, but it is the one the law actually gives us.
What to watch
- Whether states issue clearer guidance for companies supplying or supporting military AI.
- How legal reviews address dual-use data centres and other assets serving both civilian and military users.
- Whether future cases clarify how existing rules apply to technical assistance and AI-supported operations.
Discussion spark: Should states create clearer, specific rules for companies supporting military AI, or is existing international humanitarian law adequate if applied case by case?
Sources and evidence
- Source update (3 November 2026, 00:41 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.