Air Force AI teaming tests show promise in generating battle options quickly but suffer from "hallucinations," producing non-viable plans due to missed nuances. Human oversight remains crucial to ensure effective decision-making.