网络安全专家 AI 发现的漏洞 很快或将引发新的安全隐患Vulnerabilities discovered by Cybersecurity Expert AI could soon trigger new security risks.


AI free token/AI免费积分
网络安全专家 AI 发现的漏洞 很快或将引发新的安全隐患
感谢IT之家网友 愚公骑马 的线索投递!

IT之家 10 月 25 日消息,路透社采访的七名网络安全专家认为,随着 AI 工艺不断揭示软件中隐藏的安全漏洞,类似本周全球多地政府网站接连中断服务的事件,几乎肯定会再次发生。

美国马萨诸塞州网络威胁情报公司 VulnCheck 即将发布的研究表明,AI 推动了软件漏洞发现数量的增长,过去一年发现的漏洞数量同比翻了一番。

数字基础设施公司 Cloudflare 的 CEO 马修 · 普林斯指出:“我知道,其他一些关键软件中还存在不少严重漏洞。…… 我认为,接下来两年,笔者每个星期都会觉得天要塌了。”

他预计,随着 AI 陆续发现互联网基础设施中长期隐藏的安全隐患,将来几个月网络安全行业将异常忙碌。

据IT之家了解,本周,欧洲及其他地区的政府网站接连中断服务,起因是广泛应用于互联网的老牌编程语言 PHP 被发现存在安全漏洞。各网站运营者随即紧急修复软件,防止黑客大规模利用漏洞发动攻击。部分运营者担心来不及修补,索性先关闭网站,待确保安全后再恢复运行。

VulnCheck 研究员帕特里克 · 加里蒂指出,PHP 漏洞引发的紧急修复工作,让原本就不平静的网络安全行业更加忙碌。本周,PHP 替代工艺 Node.js 和流媒体软件 Plex 也被披露存在漏洞,AI 网关 LiteLLM 和开发平台 Railway 卷入安全入侵事件。

另外,印度数字证书颁发机构 eMudhra 被发现使用安全性极差的加密密钥,可能危及互联网各地的使用者。“如果一周就发生了四五起这样的事件,以后恐怕只会更多。只要平台直接连接互联网,就可能需要越来越频繁地紧急处理安全难题。”

现代软件的代码库规模庞大,程序错误往往隐藏在意想不到的地方,其中一些缺陷足以成为黑客入侵平台的突破口。

过去,发现程序漏洞需要一定的专业工艺,而将漏洞转化为可以攻破平台的攻击工具,通常还需要经验多元的工艺人员花费数周甚至数月。漫长的开发过程给了防御方喘息之机,可以利用时间差修补漏洞,逐步推出软件更新。

AI 打破了原有的时间差。AI 模型可以全天候检查软件代码,接连找出漏洞,并在短短数小时内开发出利用漏洞的攻击程序。

美国弗吉尼亚州网络安全公司 RunSafe Security 的 CTO 肖恩 · 弗赖指出,过去几个月,AI 模型已经用实际表现证明,网络安全行业必将因而发生改变。“发生根本性变化,而且毫无争议的,是时间。不仅发现漏洞所需的时间变了,攻击者将漏洞转化为攻击手段的速度也变了。”

接受路透社采访的专家一致认为,留给防御方修补漏洞的时间越来越少。但 AI 是否会导致黑客攻击泛滥,他们仍有分歧。

文件传输工具 curl 的主要开发者丹尼尔 · 斯滕伯格认为,AI 模型发现的漏洞往往比较特殊,恶意攻击者很难加以利用。

网络安全公司 ArmorCode 的资深首席应对思路工程师拉米 · 拉赫曼也指出:“披露的漏洞越来越多,并不意味着成功发动的网络攻击必然随之增加。”

本周引发全球关注的 PHP 漏洞,就给出了一些初步佐证。威胁情报公司 Hunt.io 称,与亚马逊和 Cloudflare 一样,该公司也发现有人密集尝试利用漏洞发动攻击。但 Hunt.io 研究主管埃斯特万 · 博尔赫斯指出,截至美东时间周四 14 时,从追踪结果看,网络犯罪分子仍未开发出可以实际利用该漏洞的有效工具。

Cloudflare 的普林斯则“看到过”部分平台遭到入侵的零星迹象,表明这些入侵事件与 PHP 漏洞有关。在他看来,AI 接连发现大量漏洞,虽然会给网络安全行业带来冲击,但从长远来看仍然利大于弊。

他将整个过程比作清除鬼屋里的怪物:与怪物正面交锋固然可怕,任由它们一直躲在紧闭的门后,后果却更加严重。“现实是,笔者将可以清除大量安全隐患。”

本文来源:https://www.ithome.com/1/017/592.htm
若有版权问题,请来信告知:13951431913#139.com 或在右方提交评论。

According to IT Home on October 25, seven cybersecurity experts interviewed by Reuters believe that as AI technology continues to uncover hidden software vulnerabilities, incidents similar to this week's global government website outages are almost certain to happen again.

Upcoming research from Massachusetts-based cyber threat intelligence firm VulnCheck indicates that AI has driven a growth in discovered software vulnerabilities, with the number found over the past year doubling year-over-year.

Matthew Prince, CEO of digital infrastructure company Cloudflare, pointed out: "I know that there are still many serious vulnerabilities in other critical software... I think that for the next two years, every week I will feel like the sky is falling.

He predicts that the cybersecurity industry will be exceptionally busy in the coming months as AI gradually uncovers long-hidden security vulnerabilities in internet infrastructure.

According to IT Home, government websites across Europe and other regions suffered consecutive outages this week due to a security vulnerability discovered in the widely used, long-standing programming language PHP. Website operators rushed to patch the software to prevent hackers from exploiting the flaw to launch large-scale attacks. Some operators, fearing they could not patch their systems in time, simply shut down their websites, planning to restore them only after ensuring their security.

VulnCheck researcher Patrick Garrity pointed out that urgent repairs sparked by PHP vulnerabilities have kept the already restless cybersecurity industry even busier. This week, alternative technologies Node.js and streaming software Plex also saw disclosed vulnerabilities, while AI gateway LiteLLM and development platform Railway were caught up in security breaches.

Additionally, Indian certificate authority eMudhra was found using highly insecure encryption keys, potentially compromising users across the internet. "If four or five such incidents occur in just one week, there will likely be even more in the future. As long as platforms are directly connected to the internet, they may need to handle security challenges with increasing frequency.

Modern software codebases are massive, and bugs often lurk in unexpected places, where some defects can provide just the opening hackers need to breach the platform.

In the past, discovering software vulnerabilities required significant expertise, while developing these flaws into tools capable of breaching platforms typically took experienced specialists weeks or even months. This prolonged development process gave defenders a breathing spell, allowing them to use the time gap to patch vulnerabilities and gradually roll out software updates.

AI has disrupted the traditional time lag. AI models can scan software code around the clock, continuously identify vulnerabilities, and develop exploits in a matter of hours.

Vulnerabilities discovered by Cybersecurity Expert AI could soon trigger new security risks.

Sean Frey, CTO of Virginia-based cybersecurity firm RunSafe Security, noted that over the past few months, AI models have demonstrated through their performance that the cybersecurity industry is inevitably going to change as a result. "What has changed fundamentally and indisputably is time. Not only has the time required to discover vulnerabilities changed, but so has the speed at which attackers convert vulnerabilities into exploits.

Experts interviewed by Reuters agree that defenders have increasingly less time to patch vulnerabilities. However, they remain divided on whether AI will lead to a proliferation of hacking attacks.

Daniel Stenberg, the lead developer of the file-transfer tool curl, believes that vulnerabilities discovered by AI models tend to be rather unique, making them difficult for malicious attackers to exploit.

Rami Rahman, Senior Principal Response Engineer at cybersecurity firm ArmorCode, also pointed out: "The increasing number of disclosed vulnerabilities does not necessarily mean that successful cyberattacks will increase accordingly.

The PHP vulnerability that has drawn global attention this week has offered some initial confirmation. Threat intelligence firm Hunt.io stated that, like Amazon and Cloudflare, it has observed intensive attempts to exploit the vulnerability. However, Hunt.io Research Director Esteban Borges noted that as of 2 p.m. Eastern Time on Thursday, tracking results indicated that cybercriminals have not yet developed effective tools to actually exploit the flaw.

Prince from Cloudflare has observed sporadic signs of platform intrusions linked to PHP vulnerabilities. He believes that while the continuous discovery of massive vulnerabilities by AI may disrupt the cybersecurity industry, the benefits will ultimately outweigh the drawbacks in the long run.

He likened the process to clearing monsters out of a haunted house: while confronting them head-on is terrifying, the consequences of letting them lurk behind closed doors are far worse. "The reality is, the author can eliminate a massive amount of safety hazards.

Source: https://www.ithome.com/1/017/592.htm If you have any copyright issues, please contact us at 13951431913@139.com or leave a comment on the right.

暂无评论,快来发表第一条评论吧!No comments yet. Be the first!

📮 需求咨询📮 Inquiry