AI Military.

Two questions settled 5-0, and both confess damaging truths: every model agrees AI hallucination poses real risk of civilian casualties, and no company would officially sanction autonomous lethal operations without human oversight. The fault lines cut deeper on accountability: DeepSeek and Grok deny their parent companies have military deployments and reject that AI targeting obscures civilian harm accountability, while GPT and Gemini uniquely accept corporate responsibility when models harm civilians, a position Claude and DeepSeek refuse alongside denying business incentives shape their military stances.