AI 幫助發現 PHP 漏洞，使攻擊者更容易利用

security researchers at companies like Calif, OpenAI, and Wiz used ai tools to uncover years-old vulnerabilities in the php web scripting language, known as php2shell. these vulnerabilities were known to the php team for years but were deemed too complex to exploit manually. ai models helped researchers create practical exploitations, making it easier for attackers to quickly exploit the flaws. millions of exploitation attempts were made within hours of the vulnerabilities being published, according to analysis from cloudflare. this incident highlights how ai models can magnify existing dynamics of vulnerability discovery and put pressure on under-resourced it teams, developers, and admins to keep up.