Sentinel — A Methods & Positioning Note

This note explains what Sentinel is, how it reaches its conclusions, and — just as importantly — what it is not. Sentinel is an offline decision-support instrument, not a legal adviser and not an official export-control authority. Every position it takes is dated and labelled; none of it should be relied upon without confirming the current primary sources. Any serious AI or autonomous-system programme has to satisfy two distinct bodies of law that are almost never assessed together. The first asks whether it is lawful to build and use the system — its risk tier under the EU AI Act (Regulation (EU) 2024/1689 [2024] OJ L, 2024/1689) and the obligations that tier triggers. The second asks whether it is lawful to move the system, its source, its model weights, or its underlying compute across a border — the EU Dual-Use Regulation (Regulation (EU) 2021/821 [2021] OJ L206/1) and the control lists of the Wassenaar Arrangement, one of the regimes its Annex I implements, with US EAR/ITAR exposure layered on top. In practice these are handled by different teams, different advisers, and different timelines, and the seam between them is exactly where a compliant-looking product becomes an unlawful export — or where an export-cleared item quietly breaches a prohibited-practice rule. Sentinel's premise is that the two questions draw on almost identical inputs — what the system does, its data and autonomy, its sectors, and the jurisdictions of build and use — and can therefore be answered together, in a single structured pass, early enough to change the design. Positioning note; not peer reviewed.