The chatbot cannot be held responsible.
Whoever is using chatbots for selecting targets is incompetent and should likely face war crime charges.
Has it been stated authoritatively somewhere that this was an AI-driven mistake?
There are myrid ways that mistake could have been made that don't require AI. These kinds of mistakes were certainly made by all kinds of combatants in the pre-AI era.
Yeah yeah, they probably had a human in the loop, that’s not really the point though.