Wondering: If someone sends me spam that ends up showing in my email list, will Apple on-device AI incorporate stuff there into what it does? Will it end up in my calendar, or in my phone numbers for other parties? Will this be a type of mal-whatever injection vector? Is this like errors in Wikipedia, etc., affecting LLM responses as definitive statements? How will I know it's being used?
@danb We’ve seen calendar injection spam for years (from the “find events in mail” setting).
