11-19 更新 实时
首页/ 算力/ 正文

AI 助发现互联网通用软件的严重漏洞

算力 来源:TechXplore 2026-11-19 21:40 0 阅读 机器翻译
AI 助发现互联网通用软件的严重漏洞

继续使用 Google、Facebook 或 Apple,即表示你同意使用条款。

AI 助力发现无处不在的互联网软件中的严重漏洞

作者:Doug Jacobson,The Conversation

编辑:Swati Mestri,审校:Andrew Zinin

本文已按照 Science X 的编辑流程和政策进行审核。编辑在确保内容可信度的同时,突出了以下属性:

由研究人员撰写

添加到首选来源

如果你今天查看过银行余额、预订过航班,或是在网上订购过食品杂货,那么你的请求很可能由一个用 Java 编写的计算机程序来处理。

近三十年来,Java 编程语言一直是互联网背后默默耕耘的主力之一。尽管大多数人在手机或笔记本电脑上很少直接与它打交道,但它是众多支撑现代商业、金融和政府的大型后台服务器应用的核心。

正是由于这种无处不在的普及度,美国网络安全和基础设施安全局(CISA)发布的一项新的关键安全警报才引起了全球技术团队的关注。安全研究人员在 Java 的一个核心组件中发现了一个严重漏洞,被命名为 Java4Shell。如果不及时修补,该漏洞可能让攻击者诱使存在漏洞的服务器执行恶意代码。

作为一名研究大型软件系统如何失效以及如何加以防御的网络安全研究人员,我认为这一发现既是系统管理员的一次常规行动号召,也是更广泛转变中的又一个里程碑:人工智能正在迅速改变关键软件中隐藏漏洞的发现方式。

Java 编程语言简史。图片来源:Coding Animated

要理解这个漏洞,有必要先了解 Java 最初为何如此流行。

通常,软件必须针对其运行所依赖的特定计算机硬件和操作系统进行定制。Java 借助一个名为 Java 虚拟机的巧妙软件解决了这一难题。它本质上是一个运行其他程序的程序。软件开发人员只需编写一次 Java 应用,Java 虚拟机便能在服务器操作系统之上提供一个统一的基础层,无论底层机器是什么,都能运行该应用。

除了运行应用程序之外,Java 虚拟机还提供了一套标准化的代码,用于处理日常任务,例如读取文件、管理网络连接以及格式化传入数据。

这些日常任务之一就是 Base64 编码。计算机经常需要通过网络文本协议发送复杂数据,例如图像、附件或身份验证令牌。由于原始二进制数据在传输过程中可能损坏,系统会用 Base64 将其转换为由标准文本字符组成的字符串。当 Java 服务器收到该字符串时,其 Base64 库会将其解码回原始形式,供应用程序使用。

漏洞内幕,以及 AI 为何重要

新发现的漏洞存在于 Oracle 的 Java 虚拟机实现中的 Base64 解码库。这意味着任何使用该版本 Java 虚拟机的应用程序都面临风险。

根据对该漏洞的技术分析,一段精心构造的畸形 Base64 数据可以压垮该库的处理能力,使攻击者得以突破正常边界、执行未授权命令,并有可能取得服务器的控制权。

如果这听起来似曾相识,它遵循的是近期另一个基础性网络语言 PHP 曝出高关注度漏洞时的相同模式。与 PHP 漏洞一样,这个 Java 漏洞也是由网络安全公司 Calif 使用 AI 工具搜索漏洞时发现的。

Calif 官员表示,在该 Java 漏洞公开之前,公司已通知美国、加拿大、澳大利亚、新西兰和英国政府,以及包括 Google 和 Microsoft 在内的主要科技公司。随后,网络安全和基础设施安全局发布了警报,敦促各机构立即修补该漏洞。

过去,要从数百万行成熟代码的深处找出隐蔽的内存错误或解析错误,需要人类专家耗费数月进行艰苦的人工审查。如今,安全研究人员正将 AI 工具与传统代码分析相结合,系统性地排查软件库中那些人类可能多年都忽略的边界情况。

每当一个关键漏洞影响到某个基础软件时,新闻标题往往会显得危言耸听。实际上,即便需要迅速采取行动,也有若干理由让我们保持客观判断。

首先,好消息是,一位负责任的安全研究人员发现了该漏洞并进行了私下报告,让软件维护者有时间在攻击者大规模利用之前开发并发布修复程序。

其次,普通用户不必为自己的个人设备恐慌。这是影响企业应用和云基础设施的问题,而非家用笔记本电脑、智能手机或平板电脑。

第三,即使在存在漏洞的服务器上,在现实环境中利用这个漏洞也并非总是易事。大多数现代 Web 系统不会把原始的 Java 解码器直接暴露在公开互联网上。相反,传入流量通常会经过防火墙、负载均衡器和 Web 代理等多种软件,这些软件会在数据到达底层 Java 应用之前对其进行检查、清洗或拒绝畸形数据。

尽管如此,“更难利用”并不意味着不可能,尤其是对于内部系统或会把用户提供的、被称为 Token 的信息片段直接传给 Java 解码器的定制应用而言。与任何基础架构的修复一样,最好的防御很直接:及时应用最新的安全更新。

随着 AI 加速发现长期隐藏的软件漏洞,保持数字基础设施及时打补丁仍是最有效的网络安全实践之一。

由 The Conversation 提供

这篇报道的作者是谁?

Swati Mestri 拥有电子工程学士学位,自 2019 年起从事内容编辑工作。她曾编辑技术、医疗健康与材料科学领域的研究文档,对技术与太空领域尤为关注。完整简介 →

物理学硕士,具备科研经历。资深科学新闻爱好者。在 Science X 的编辑工作中发挥关键作用,助力其取得成功。完整简介 →

本文依据 Creative Commons 许可协议转载自 The Conversation。阅读原文。

引用:AI helps uncover critical flaw in ubiquitous internet software(2026 年 11 月 19 日),2026 年 11 月 19 日检索自 https://techxplore.com/news/2026-11-ai-uncover-critical-flaw-ubiquitous.html

本文件受版权保护。除出于私人学习或研究目的的合理使用外,未经书面许可不得以任何形式复制任何部分。内容仅供参考。

随着运营方着手处理一个由 AI 发现的网页编程语言漏洞,部分网站已下线

各方正紧急修补一个可能影响全球 70% 网站的软件漏洞

一个影响所有网站 70% 的严重漏洞让网络安全界陷入恐慌

关于这个由 AI 发现、波及互联网大片区域的安全漏洞,你需要了解什么

“技术债到期”:AI 发现的漏洞预示着混乱将至

莲茎与生物膜利用蒸发的水分发电,为小型电子设备供电

迄今最快的软体机器人无需电子元件即可运行

轻若羽毛的飞行机器人自带传感器和计算机

微观“聚合物纤维森林”有助于为电子设备散热

结构电池电极使能量存储能够置于弯曲的无人机支腿内部

迷迭香与魔方:一种气味如何帮助我们在混合现实中记住所学内容

获取即时摘要文本(Gist)

AI 辅助分析发现 Oracle 的 Java 虚拟机 Base64 解码器中存在一个缺陷,可能允许通过畸形输入执行恶意代码。是否可被利用取决于应用暴露面和输入处理方式。建议对受影响的服务器及时打补丁。

本摘要由 LLM 自动生成。完整免责声明

译文由英文原文机器翻译,术语按行业惯例统一,可在文末展开原文核对。

查看英文原文(AI helps uncover critical flaw in ubiquitous internet software)

By continuing with Google, Facebook or Apple, you agree to the Terms of Use.

AI helps uncover critical flaw in ubiquitous internet software

by Doug Jacobson, The Conversation

edited by Swati Mestri, reviewed by Andrew Zinin

This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility:

written by researcher(s)

Add to Preferred Sources

Java code Credit: Unsplash/CC0 Public Domain

If you checked your bank balance, booked a flight or ordered groceries online today, there is a good chance your request was handled by a computer program written in Java.

For nearly three decades, the Java programming language has been one of the quiet workhorses of the internet. While most people rarely interact with it directly on their phones or laptops, it sits at the heart of most large, behind-the-scenes server applications that power modern business, finance and government.

That ubiquity is why a new critical security alert from the U.S. Cybersecurity and Infrastructure Security Agency drew the attention of technology teams worldwide. Security researchers uncovered a severe flaw, dubbed Java4Shell, in a core Java component. If left unpatched, the bug could allow an attacker to trick a vulnerable server into executing malicious code.

As a cybersecurity researcher who studies how large-scale software systems fail and how to defend them, I see this discovery as both a standard call to action for system administrators and another milestone in a broader shift: Artificial intelligence is rapidly changing how hidden flaws in critical software are discovered.

A brief history of the Java programming language. Credit: Coding Animated

To understand the vulnerability, it helps to know why Java became so popular in the first place.

Normally, software must be tailored to the specific computer hardware and operating system it runs on. Java solved that headache with a clever piece of software called the Java Virtual Machine. It is essentially a program that runs other programs. Software developers write their Java application once, and the Java Virtual Machine provides a uniform foundation on top of a server's operating system, running that application regardless of the underlying machine.

In addition to running the application, the Java Virtual Machine provides a standardized set of code for everyday tasks, such as reading files, managing network connections and formatting incoming data.

One of those routine tasks is Base64 encoding. Computers often need to send complex data, such as images, attachments or authentication tokens, over text-based internet protocols. Because raw binary data can be corrupted in transit, systems convert it into a string of standard text characters using Base64. When a Java server receives that string, its Base64 library decodes it back into its original form so the application can use it.

Inside the flaw and why AI matters

The newly discovered vulnerability resides in the Base64 decoding library in Oracle's implementation of the Java Virtual Machine. This means that any application using this version of the Java Virtual Machine is at risk.

According to technical analysis of the flaw, a carefully crafted piece of malformed Base64 data can overwhelm the library's processing, allowing an attacker to break out of normal boundaries, execute unauthorized commands and potentially gain control of the server.

If this sounds familiar, it follows a pattern seen recently with a high-profile vulnerability in PHP, another foundational web language. Like the PHP vulnerability, this Java flaw was uncovered by cybersecurity firm Calif using AI tools to search for vulnerabilities.

Calif officials said the company alerted the governments of the United States, Canada, Australia, New Zealand and Britain, and major tech companies, including Google and Microsoft, before the Java vulnerability became public. The Cybersecurity and Infrastructure Security Agency then issued its alert, urging organizations to immediately patch the flaw.

Historically, identifying subtle memory or parsing errors buried deep within millions of lines of mature code required months of painstaking manual review by human experts. Today, security researchers are pairing AI tools with traditional code analysis to systematically probe software libraries for edge cases that humans might overlook for years.

Whenever a critical vulnerability affects a foundational piece of software, headlines can sound alarmist. In reality, there are several reasons to maintain perspective while still acting quickly.

First, the good news is that a responsible security researcher discovered and privately reported this flaw, giving people who maintain software time to develop and release a fix before attackers could widely exploit it.

Second, everyday users do not need to panic about their personal devices. This is an issue affecting enterprise applications and cloud infrastructure, not home laptops, smartphones or tablets.

Third, even on vulnerable servers, exploiting this bug in the real world is not always straightforward. Most modern web systems do not expose raw Java decoders directly to the open internet. Instead, incoming traffic typically passes through several types of software—firewalls, load balancers and web proxies—that inspect, sanitize or reject malformed data before it reaches the underlying Java application.

Still, "harder to exploit" does not mean impossible, especially for internal systems or custom applications that pass user-supplied bits of information, called tokens, directly to Java's decoder. As with any foundational infrastructure repair, the best defense is straightforward: promptly applying the latest security updates.

As AI accelerates the discovery of long-hidden software flaws, keeping the digital infrastructure patched remains one of the most effective cybersecurity practices.

Provided by The Conversation

Who's behind this story?

Swati Mestri

Swati Mestri holds a bachelor's degree in Electronics Engineering and has worked as a content editor since 2019. She has experience editing research documents across technology, health care, and materials science, and has a particular interest in technology and space. Full profile →

Master's in physics with research experience. Long-time science news enthusiast. Plays key role in Science X's editorial success. Full profile →

This article is republished from The Conversation under a Creative Commons license. Read the original article.The Conversation

Citation: AI helps uncover critical flaw in ubiquitous internet software (2026, November 19) retrieved 19 November 2026 from https://techxplore.com/news/2026-11-ai-uncover-critical-flaw-ubiquitous.html

This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.

Some websites go dark as operators address an AI-detected flaw in a web programming language

Rush is on to patch a software vulnerability that could impact 70% of websites worldwide

A critical flaw affecting 70% of all websites has the cybersecurity world panicked

What to know about the AI-detected security bug impacting wide swaths of the internet

'Tech debt coming due': AI-found flaw signals chaos ahead

Lotus stems and biofilms generate electricity from evaporating water to power small electronics

The fastest soft robot ever built doesn't need electronics to function

Featherweight flying robot carries its own sensors and computers

Microscopic 'polymer fiber forest' helps keep electronics cool

Structural battery electrode enables energy storage inside curved drone legs

Rosemary and Rubik's Cubes: How a scent helps us remember what we learn in mixed reality

Get Instant Summarized Text (Gist)

AI-assisted analysis identified a flaw in Oracle’s Java Virtual Machine Base64 decoder that could allow malicious code execution through malformed input. Exploitation depends on application exposure and input handling. Prompt patching is recommended for affected servers.

This summary was automatically generated using LLM. Full disclaimer

This site uses cookies to assist with navigation, analyse your use of our services, collect data for ads personalisation and provide content from third parties. By using our site, you acknowledge that you have read and understand our Privacy Policy and Terms of Use.

原文为英文,正文由自动抓取并保留原文。

本文由采集器自动同步

原文发布于 TechXplore(AI helps uncover critical flaw in ubiquitous internet software),标题、摘要与正文由机器翻译,术语按行业惯例统一,未经人工逐句校订。

阅读原文 ↗