1 in every 28 GenAI prompts posed a high risk of sensitive data leakage in March 2026 and 91% of organizations using GenAI tools regularly were impacted by this risk.