68% of Organizations Experienced Data Leakage From Employee AI Usage | Security Magazine
A Metomic survey reveals significant AI-related data security challenges despite organizational confidence. While 90% of respondents expressed confidence in security measures and 91% believed training initiatives were successful, over half reported regular malware attacks, phishing, and breaches linked to improper AI implementation. The study found 68% of organizations experienced data leakage from employees sharing sensitive information with AI tools, yet only 23% have implemented comprehensive AI security policies. Security leaders identified fostering strong security culture as their primary 2025 challenge, with 44% now prioritizing security infrastructure oversight—largely focused on securing AI systems. This represents a shift from previous emphasis on security operations. Ransomware has overtaken phishing as the top U.S. security concern, with AI-enabled attacks showing increased sophistication, whilst UK organizations report heightened third-party supplier risks driven by AI solution integration.