Nearly half (45%) of organisations opted not to report an AI-related security breach due to concerns over reputational damage.