The issue
Last week, two security developments highlighted how the speed and scale of AI cyberattacks are expanding the attack surface and impacting cyber defenders.
PHP Vulnerability
PHP is the open-source programming language behind a large share of the world’s websites, including those built on WordPress. On the morning of Oct. 21, PHP and Wiz Research disclosed two critical flaws in the language’s core PHP code (CVE-2026-109618 and CVE-2026-109619). According to Wiz, AI helped find and fix vulnerabilities that had gone undetected in the PHP codebase for more than a decade.
Together the two flaws let an attacker take control of an affected server without logging in and read sensitive data from its memory. The flaws were rated 9.2 out of 10 on the Common Vulnerability Scoring System (CVSS), which is classified as critical. The flaws affect PHP 5.4 through 8.5.11. As a result, organizations should update to 8.5.12, 8.4.27, 8.3.36 or 8.2.35. PHP versions 8.1 and earlier are no longer supported and won’t be patched. A tracker from Censys shows patching progress across internet-facing PHP services. Google and other threat teams are tracking related activity, and defenders should treat the flaws as under active exploitation.
ConnectWise incident
That same afternoon, Microsoft reported that a China-linked hacking group recently used ARTEX, an open-source AI penetration testing framework, to automate attacks against organizations worldwide through exposed ConnectWise credentials. ConnectWise makes software that IT service providers use to manage and back up their customers’ systems. Separate reporting from Huntress Labs found that attackers had compromised the digital keys used to connect ConnectWise Cloud Backup (formerly SkyKick) to customers’ Microsoft 365 accounts and were using them to gain unauthorized access.
The incident led ConnectWise to temporarily suspend certain cloud-based services. According to Microsoft, the attackers first took control of the identities that ConnectWise’s cloud apps use to access customers’ systems. AI agents operating on behalf of the threat actor then used those stolen credentials to log in to customer environments. The agents exploited the trust that customers had already granted to the ConnectWise apps.
Implications for organizations
These incidents reinforce the importance of defending at machine speed. To that end, our latest brief lays out ten crucial imperatives for security leaders today. This paper offers a framework to help companies address the emerging threat of AI-enabled adversaries capable of identifying and exploiting a wide range of cyber vulnerabilities at scale. Read Defending at Machine Speed.
Read more about AI-enabled cyberattacks. Visit our Cybersecurity page.




