44% of AI-first organizations report that AI was directly exploited in their most recent security incident, compared to 6% of non-AI-first organizations.