The Military AI Industry’s Involvement in Armed Conflict: Implications Under International Humanitarian Law – Analysis

Eurasia Review | Laura Bruun, Marta Bo

Iranian strikes on Amazon Web Services data centres in Bahrain and the United Arab Emirates highlight the expanding operational role of private technology contractors in active warzones. Dual-use infrastructure supports military targeting tools alongside civilian workloads across conflicts in Gaza, Ukraine, and Venezuela. This integration complicates civilian protection standards established under Additional Protocol I to the Geneva Conventions.

Commercial assets lose legal immunity under Article 52 when their neutralization yields a definite military advantage to opposing forces. Forward-deployed engineers embedded with military units risk losing civilian protection under Article 51 if their technical support constitutes direct participation in hostilities. Corporate contractors and individual software developers face secondary criminal liability under Article 25 of the Rome Statute for aiding and abetting unlawful attacks. Legal boundaries remain contentious. Under Common Article 1, states retain primary legal responsibility to regulate private contractors and conduct mandatory weapon reviews under Article 36 before operational deployment.

Comment

Integrating commercial algorithms into command nodes shifts operational decision speed from uniformed officers to software engineers. Systems like the Maven Smart System rely on private contractors to maintain real-time data fusion pipelines during combat operations. This arrangement creates command ambiguities when algorithm updates alter target prioritization parameters mid-mission. Uniformed commanders risk delegating target validation to contractor algorithms without explicit legal authorization.

The friction centers on latency between algorithm recalibration and command approval. In US Central Command operations, forward-deployed technical teams push cloud-based model revisions directly into sensor-to-shooter loops. This continuous integration pipeline bypasses standard military air tasking order verification cycles. Consequently, CENTCOM air tasking orders risk processing targets derived from contractor patches that bypass Combined Air Operations Center review procedures.

Strategic Question for Discussion

If forward-deployed contractors continually update the Maven Smart System during active operations, which component retains ultimate legal responsibility for target selection—the Combined Air Operations Center commander or the software vendor?

The operational pattern indicates that legal responsibility remains anchored to the uniformed commander who authorizes the strike, regardless of automated data inputs. However, reliance on real-time contractor model updates at Combined Air Operations Centers blurs factual causation when algorithms misidentify civilian targets. My assessment is that military tribunals will increasingly struggle to isolate commander intent from proprietary code flaws during war-crimes investigations.

Share your assessment in the comments below.

To read in details click here