AI-assisted bug hunting is reportedly speeding up defensive patching and exploit development
AI securityUnverifiedUpdated:
Reporting on PHP2Shell describes AI assistance speeding vulnerability discovery and exploit development. It also covers defensive patching and service suspensions, while the eventual impact remains uncertain.
PHP2Shell attempts and pressure on patching time
404 Media reported that researchers from Wiz, Calif, and OpenAI found the PHP2Shell vulnerabilities using AI. It also described thousands of attempted exploits on the Cloudflare network before public disclosure, escalating after disclosure, while organizations patched systems or temporarily removed some services from operation. The report said there had not been immediate, widespread exploitation, and that some low-effort exploits produced with large language models did not work.
The report provides no comparable measurement of vulnerability discovery or exploit development before AI-assisted tools. It describes coordinated disclosure and patching as the preferred security practice.
According to the report, AI-assisted researchers and attackers are speeding up vulnerability discovery and exploit development. The PHP2Shell disclosures illustrated both rapid exploit attempts and a shortened period for defenders to patch, though the ultimate impact remains unclear.
The reported events show the same AI-assisted capabilities being used to find flaws, prepare fixes, and develop offensive tools. The report emphasizes that the speed of deploying patches across individual organizations remains uncertain.
Sources
- 404 MediaAI Is Changing Everything About Hacking. No One Knows What Happens NextRead source →