Security briefing: February 2026

In February 2026, cybersecurity was dominated by issues related to artificial intelligence (AI), highlighting both its potential and the security risks it poses. AI-driven attacks, such as AI recommendation poisoning and AI-assisted ransomware, underscored the vulnerabilities in systems due to weak credentials and exposed management ports. Significant incidents included the rapid exploitation of the BeyondTrust vulnerability (CVE-2026-1731) and the compromise of Fortinet FortiGate devices across 55 countries. The month also saw AI being used to compress attack kill chains, as demonstrated by the Sysdig Threat Research Team, which showed how attackers could gain admin privileges in just eight minutes. Meanwhile, the integrity of AI systems' training data was called into question due to hidden manipulations by legitimate organizations. Despite the rapid evolution of AI threats, the importance of traditional security measures such as credential management, patching, and identity hygiene remains crucial, as emphasized by the need for real-time detection over reactive measures. The European Commission and the French Ministry of Finance dealt with separate cyber incidents, while companies like Anthropic and Sysdig introduced AI tools for security enhancement, even as the misuse of AI continues to grow.