This paper investigates how varying levels of AI disclosure in news articles impact readers' trust and behavior. It reveals that only detailed disclosures significantly reduce trust, while all forms of disclosure increase source-checking, highlighting a complex trade-off between transparency and public trust.