AI for Military Support
A study of 2,015 Israeli military personnel found that soldiers exhibited "algorithmic aversion" to AI decision-support systems, particularly in high-collateral-damage scenarios, contrary to fears of automation bias; however, integrating "explainable AI" features reduced this aversion and improved evaluation of recommendations.
A study with 2,015 Israeli military personnel found that soldiers exhibit “algorithmic aversion” to AI decision-support systems in targeting, especially when collateral damage is high—contrary to fears of automation bias. However, integrating “explainable AI” features reduced this aversion and encouraged more thoughtful evaluation of recommendations. The research, using a high-fidelity replica of a real military AI system, shows trust in battlefield AI is dynamic, shaped by individual predispositions, operational stakes, and interface design. The findings underscore human agency’s enduring role in high-stakes military decisions.