I'm not sure what's better/worse: 1. A LLM did the targeting 2. Targets were selected and evaluated by humans and the many-years old change of building shape

I'm not sure what's better/worse: 1. A LLM did the targeting 2. Targets were selected and evaluated by humans and the many-years old change of building shape and purpose didn't get caught during target eval.