Details
- OpenAI said its models helped researchers discover two critical security flaws in PHP through its Patch the Planet initiative.
- The vulnerabilities were found during a two-month coordinated effort involving PHP maintainers, researchers, and security providers.
- Calif identified one critical flaw using OpenAI’s GPT-5.6 Sol model, while Wiz independently discovered and reported both vulnerabilities.
- The flaws, tracked as CVE-2026-109618 and CVE-2026-109619, could enable unauthenticated remote code execution on vulnerable PHP deployments.
- PHP maintainers led remediation and released emergency updates, including PHP 8.5.12, 8.4.27, 8.3.36, and 8.2.35.
- OpenAI, Wiz, Cloudflare, and other partners contributed technical research, protection measures, and disclosure coordination.
- PHP runs an estimated 70% of websites, giving the discovery broad potential significance for web infrastructure.
- The initiative pairs AI-assisted vulnerability research with expert review and coordinated patching before public disclosure.
Impact
The announcement positions AI-assisted code auditing as a practical defensive capability rather than merely a research demonstration. Finding vulnerabilities before disclosure can extend defenders’ preparation time, but the operational value depends on human validation, responsible coordination, and rapid patch adoption. Because PHP underpins a large portion of the web, the episode could encourage maintainers and security teams to use capable models for audits while increasing pressure to establish safeguards against misuse.