PHP编程语言中发现的两个关键安全漏洞已隐藏超过十年，由人工智能工具首次揭露，随即引发全球网络安全危机12。这些漏洞影响约占全球互联网70%的网站，可能导致未认证的远程代码执行，使攻击者能够控制服务器并盗取数据23。Wiz公司将其称为"AI检测到的有史以来最严重的漏洞"，而OpenAI则评价其为"近期最具影响力的安全问题之一"12。
漏洞由Wiz和Calif公司的研究人员使用AI工具在数周内发现，并可追溯至2004年的PHP 5版本12。PHP开发团队于周三下午发布了修复补丁1。然而，漏洞详情在周二提前于社交媒体泄露，导致黑客在官方补丁发布前数天内已开始利用这些漏洞进行攻击34。Amazon在10月17日和10月21日遭到多次攻击尝试，日本银行和航空公司网站也因此离线34。
全球政府和网络安全机构随即发布紧急预警。美国国家安全局警告称该漏洞"通常提供直接通往服务器文件系统或数据库的路径"，并敦促无法立即修补的组织下线系统13。英国国家网络安全中心、欧盟、澳大利亚等机构均发布了类似警报1。日本首相高市早苗在社交媒体平台X上呼吁立即采取行动，并敦促检查网站是否运行PHP12。白宫表示联邦机构正与AI实验室合作加快修复进度2。
维基百科、芬兰政府、BBC、爱丁堡市议会等多个机构已主动下线部分系统以防受攻击12。澳大利亚AI战略委员会首席执行官Anna-Maria Arabia表示："过去需要数月才能暴露漏洞，现在只需数天"，反映出AI技术在加快漏洞发现和利用速度方面的双重影响3。
Researchers using artificial intelligence tools discovered two severe security vulnerabilities in PHP, the programming language that powers approximately 70% of the internet.23 The flaws, which had remained undetected for over a decade, were identified by teams from Wiz and Calif companies working with OpenAI.12 OpenAI described the discovery as "one of the most impactful security issues in recent years,"1 while Wiz characterized it as "the most severe vulnerability detected by AI to date."2 The vulnerabilities originated in PHP 5, dating back to 2004, and could allow attackers to execute unauthorized remote code and access server systems without authentication.23
The revelation sparked an unprecedented global response from government agencies and cybersecurity organizations.1 The U.S. National Security Agency warned that the flaws "typically provide direct paths to server file systems or databases,"3 and urged organizations unable to patch immediately to take systems offline.12 The British National Cyber Security Centre, the European Union, and Australian authorities issued similar alerts.1 Japan's Prime Minister Takayama Sanae called for immediate action on X, urging verification of website PHP usage.1 PHP developers released patches on Wednesday, and multiple high-profile organizations including Wikimedia, Finland's government, Edinburgh Council, and the BBC preemptively took systems offline as a precaution.12
Attack activity emerged before patches were fully deployed.3 Attackers targeted Amazon on October 17 and October 21,3 while critical infrastructure in Japan including banks and airlines went offline due to the vulnerabilities.4 Vulnerability details leaked on social media Tuesday, preceding the official lifting of embargo restrictions.3 Australia's AI Strategy Commission Chief Executive Anna-Maria Arabia noted the acceleration of threat timelines, stating: "What used to take months to expose in vulnerabilities now takes days."3 The White House confirmed that federal agencies were collaborating with AI laboratories to accelerate remediation efforts.2
评论
还没有评论，欢迎留下第一条。