I don't know this for certain, but I imagine there's some kind of kv store between queries and AI overviews. Maybe they could update certain overviews or redo them with a better model.
This seems more like a model-specific issue, where it’s consistently generating flawed output every time the cache gets invalid. If that’s the case, there’s not much Google can do on a case-by-case level, but we should see improvements over time as the model gets incrementally better / it becomes more financially viable to run better models at this scale.