AI Targeting and the Rendulic Rule: The Battle Over Military Liability

Military legal scholars and defense experts are analyzing how the 'Rendulic Rule' applies to the integration of artificial intelligence in combat targeting. As modern warfare shifts toward machine-speed decision-making, the debate centers on whether commanders can be held criminally liable for erroneous strikes based on AI outputs.

The Rendulic Rule originates from a 1948 Nuremberg tribunal case known as the Hostage case. In that instance, General Lothar Rendulic was acquitted of charges regarding his scorched-earth tactics because the tribunal judged his decisions based on the intelligence available to him at the time, rather than the actual outcome. In the modern era, this standard suggests that commanders may be shielded from liability if they rely on AI-augmented targeting systems to process vast amounts of Intelligence, Surveillance, and Reconnaissance (ISR) data, provided that such reliance is deemed reasonable.

The Data Deluge of Modern Combat

This legal tension arises from the shift toward Multi-Domain Operations (MDO), which require the seamless integration of capabilities across land, air, sea, space, and cyberspace. This operational shift has created a 'data deluge,' where the volume and speed of incoming information make human-only processing effectively impossible. To maintain a competitive edge, militaries are adopting AI-assisted targeting to identify and engage threats at speeds that far exceed human cognitive limits.

The Accountability Gap

As AI takes a more prominent role in the 'kill chain,' the legal framework for accountability must evolve to prevent a dangerous liability gap. Erroneous combat decisions must be both 'subjectively honest and objectively reasonable.' If the Rendulic Rule is applied too broadly, there is a risk that commanders will not be held responsible for systemic AI errors. Conversely, if the rule is ignored, commanders could be unfairly penalized for honest mistakes made under the extreme pressure of machine-speed warfare.

The Limits of Legal Precedent

Some scholars argue that existing legal frameworks are insufficient for this transition. Researchers highlight the 'law of the horse'—a concept coined by Judge Frank H. Easterbrook suggesting that general legal principles are sufficient for new technologies. However, critics argue this approach may be inadequate for the specific nuances of AI in military targeting, where the complexity of the algorithm may obscure the 'reasonableness' of a commander's reliance on it.

What remains to be seen is how international tribunals and military courts will define 'objective reasonableness' when the evidence is buried in a black-box algorithm. The coming years will likely determine whether the Rendulic Rule remains a shield for honest error or becomes a loophole for algorithmic failure.