Pharmaceutical Artificial Intelligence (AI) is reshaping pharmacovigilance, promising faster detection of safety signals hidden within mountains of adverse event (AE) data. However, as the US Food and Drug Administration (FDA) sharpens its focus on transparency, credibility, and real-world monitoring, the days of relying on opaque algorithms are numbered. In this article, we explore why sponsors and safety teams must rethink their approach to AE prediction to ensure the very tools designed to protect patients don’t become risks themselves. 2 July 2025