There is a lot of detail in this report, and why on earth are these untrustworthy AI systems being allowed access to classified material? And you have to wonder if AI hallucinations are responsible for killing on errant conclusions, inadequately analyzed by humans?