Vulnerabilities identified with the assistance of artificial intelligence are significantly more likely to result in remote code execution (RCE), as both disclosures and exploitation saw an acceleration in 2026. Research published September 30 by Google Threat Intelligence Group (GTIG) found that 50% of vulnerabilities they identified as likely AI-discovered led to RCE, compared to 26% of other CVEs, with further coverage provided by Infosecurity Magazine.
Vulnerability disclosures doubled in 2026, and exploited vulnerabilities increased from an average of 10.5 per month in 2025 to 18 per month in 2026. While zero-day exploitation saw a marginal rise, the overall growth is attributed to the rapid weaponization of known vulnerabilities, potentially aided by AI tools analyzing patches and proof-of-concept code. Between January and August 2026, medium-risk flaws constituted 58% of likely AI-discovered vulnerabilities, versus 28% of those not attributed to AI.
Google suggests this distribution reflects researchers targeting critical infrastructure with autonomous agents. Exploitation remains concentrated at the perimeter, with edge and security appliances making up 14% of exploited vulnerabilities in 2026, over 65% of which were rated high or critical risk.
Source: Infosecurity Magazine
