Trend Micro’s latest report reveals alarming trends in AI security, with vulnerabilities hitting unprecedented levels. According to the TrendAI State of AI Security Report, a staggering rise in AI-related weaknesses is noted, with many being classified as high or critical severity. The report underscores a troubling gap between the rapid deployment of AI technologies and the maturity of security practices surrounding them.
The findings indicate that almost half of the critical vulnerabilities are emerging in less mature sectors, creating significant risks. Most notably, new vulnerabilities have arisen in MCP servers, recording 95 new Common Vulnerabilities and Exposures (CVEs) in 2025 compared to virtually none previously. There’s also a dramatic surge in weaknesses found in autonomous AI systems, with reported flaws soaring by a staggering 255% in just one year. TrendAI analysts predict that 2026 could see between 2,800 and 3,600 new AI-related vulnerabilities, marking a potential 69% increase from 2025 if proactive measures are not taken. Security expert Richard Werner emphasizes the need for businesses to integrate robust cybersecurity measures in AI projects to mitigate risks effectively.
