Google Threat Intelligence Group experts report that AI is enabling cybercriminals to identify and exploit software vulnerabilities with increased speed and efficiency. The focus of these attacks is shifting away from zero-day vulnerabilities toward the rapid weaponization of already known security flaws.
Data for 2026 shows a significant surge in vulnerability discovery. In January, 5,045 errors were identified, while by August, this number reached 10,740. The frequency of exploited vulnerabilities in real-world attacks rose from 10.5 per month in 2025 to 18 per month in 2026.
AI also assists security professionals in prioritizing threats. Among vulnerabilities identified by AI, 50 percent allow for remote code execution, compared to 26 percent in the general pool of vulnerabilities. Experts note that AI helps filter out low-risk issues to focus on critical infrastructure.
Attackers are now weaponizing publicly disclosed vulnerabilities within days, utilizing them for privilege escalation, data theft, and malware deployment. Google recommends that organizations move away from mass patching and adopt threat intelligence-based vulnerability management.
