34
Artificial intelligence (AI) is accelerating the discovery of software vulnerabilities, increasing pressure on businesses and governments to
secure their systems before attackers can exploit weaknesses, according to a Reuters report published on October 23.
The report cited seven cybersecurity professionals who warned that AI- assisted vulnerability discovery could lead to more frequent cyber disruptions. Research from threat intelligence firm VulnCheck indicates that the number of disclosed software vulnerabilities has doubled year on year, with AI contributing to the increase.
The warning follows the discovery of two critical vulnerabilities in PHP, a widely used programming language that powers websites and web applications. Tracked as CVE-2026-109618 and CVE-2026-109619 and collectively referred to as PHP2Shell, the flaws can allow unauthenticated attackers to execute code remotely on vulnerable servers.
The PHP development team released security updates on October 21, addressing the vulnerabilities in PHP versions 8.5.12, 8.4.27, 8.3.36 and 8.2.35. The European Union Agency for Cybersecurity and CERT-EU have urged organisations to identify affected systems and apply the relevant updates promptly.
The vulnerabilities have triggered emergency responses, with some operators taking services offline while assessing their exposure. Reuters reported that affected PHP systems contributed to outages involving government websites in Europe and elsewhere.
The extent of successful exploitation remains unclear. Amazon Web Services (AWS) detected attempted exploitation of the PHP2Shell vulnerabilities dating back to October 17, before the security updates were publicly released. However, AWS said its investigation had not established whether those attempts would have succeeded against real-world systems. Security researchers have also reported differing findings across their monitoring environments.
AI tools can help researchers analyse large codebases, identify programming errors and investigate potential attack paths. However, the same capabilities could allow malicious actors to examine newly disclosed vulnerabilities and develop exploitation techniques more quickly, narrowing the window available for security teams to respond.
Cloudflare Chief Executive Matthew Prince told Reuters that other serious vulnerabilities remained in critical software. He warned of sustained cybersecurity pressure as AI tools uncover weaknesses that may have remained undetected for years.
The report also highlighted vulnerabilities involving Node.js and Plex, alongside security incidents affecting AI gateway LiteLLM and software development platform Railway, illustrating the broader challenges facing technology operators.
Security teams are being urged to identify exposed systems, apply vendor-issued patches and investigate suspicious activity. Organisations running PHP should verify their installed versions, follow official security guidance and prioritise internet-facing applications. Unsupported software requires particular attention because standard security updates may no longer be available.
Although AI-assisted vulnerability discovery could strengthen cybersecurity by helping developers identify and fix flaws earlier, the benefits depend on timely disclosure, effective patch management and organisations' ability to respond quickly.
As AI capabilities advance, the central challenge for defenders will be to remediate vulnerabilities before attackers can exploit them to disrupt services or compromise sensitive systems.
The report cited seven cybersecurity professionals who warned that AI- assisted vulnerability discovery could lead to more frequent cyber disruptions. Research from threat intelligence firm VulnCheck indicates that the number of disclosed software vulnerabilities has doubled year on year, with AI contributing to the increase.
The warning follows the discovery of two critical vulnerabilities in PHP, a widely used programming language that powers websites and web applications. Tracked as CVE-2026-109618 and CVE-2026-109619 and collectively referred to as PHP2Shell, the flaws can allow unauthenticated attackers to execute code remotely on vulnerable servers.
The PHP development team released security updates on October 21, addressing the vulnerabilities in PHP versions 8.5.12, 8.4.27, 8.3.36 and 8.2.35. The European Union Agency for Cybersecurity and CERT-EU have urged organisations to identify affected systems and apply the relevant updates promptly.
The vulnerabilities have triggered emergency responses, with some operators taking services offline while assessing their exposure. Reuters reported that affected PHP systems contributed to outages involving government websites in Europe and elsewhere.
The extent of successful exploitation remains unclear. Amazon Web Services (AWS) detected attempted exploitation of the PHP2Shell vulnerabilities dating back to October 17, before the security updates were publicly released. However, AWS said its investigation had not established whether those attempts would have succeeded against real-world systems. Security researchers have also reported differing findings across their monitoring environments.
AI tools can help researchers analyse large codebases, identify programming errors and investigate potential attack paths. However, the same capabilities could allow malicious actors to examine newly disclosed vulnerabilities and develop exploitation techniques more quickly, narrowing the window available for security teams to respond.
Cloudflare Chief Executive Matthew Prince told Reuters that other serious vulnerabilities remained in critical software. He warned of sustained cybersecurity pressure as AI tools uncover weaknesses that may have remained undetected for years.
The report also highlighted vulnerabilities involving Node.js and Plex, alongside security incidents affecting AI gateway LiteLLM and software development platform Railway, illustrating the broader challenges facing technology operators.
Security teams are being urged to identify exposed systems, apply vendor-issued patches and investigate suspicious activity. Organisations running PHP should verify their installed versions, follow official security guidance and prioritise internet-facing applications. Unsupported software requires particular attention because standard security updates may no longer be available.
Although AI-assisted vulnerability discovery could strengthen cybersecurity by helping developers identify and fix flaws earlier, the benefits depend on timely disclosure, effective patch management and organisations' ability to respond quickly.
As AI capabilities advance, the central challenge for defenders will be to remediate vulnerabilities before attackers can exploit them to disrupt services or compromise sensitive systems.
