11-19 更新 实时
首页/ 算力/ 正文

技术故障频发,AI 令风险加剧,这会成新常态吗

算力 来源:TechXplore 2026-11-19 18:40 0 阅读 机器翻译
技术故障频发,AI 令风险加剧,这会成新常态吗

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

技术故障不断累积,而 AI 正在推高风险。这会成为新常态吗?

作者:Wyatte Grantham-Philips、Kelvin Chan 和 Kaitlyn Huamani

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

添加到首选来源

2026 年 11 月 16 日,星期一,一名女子走过圣迭戈国际机场的行李区。图片来源:AP Photo/Gregory Bull

无法访问的电子邮箱。银行业务延迟。还有一场通信系统崩溃导致飞机停飞数小时,随后又引发了一波又一波的出行中断。

在一个日益互联的世界里,一些人担心这类中断可能成为新常态。

仅在过去一周里,微软、摩根大通和苹果,甚至美国政府的联邦航空管理局和国家安全局,都出现了看似由例行更新、安全补丁或其他故障引发的问题。特朗普政府表示,有"零"迹象表明最新的宕机事件相互关联,或由网络犯罪分子之类的恶意行为者造成。

人工智能时代也整体上提高了互联网安全的利害关系。随着 AI 工具发现新的漏洞,许多其他公司和机构正争相在漏洞被黑客盯上之前打好补丁,而这场猫鼠游戏可能引发另一类故障。

社交媒体平台上充斥着对一连串宕机事件的各种反应——烦恼、困惑、阴谋论和玩笑交织在一起。

"看到这么多故障如此密集地接连发生,我想对很多人来说,这让他们意识到,自己的生活中有多少部分已经交到了这些我们至今仍未完全理解的系统手里,"经常在 Instagram 和 YouTube 等平台发布政治相关视频的内容创作者杰克·史密斯(Jake Smith)在接受采访时说。"看到这一切在某种程度上同时发生,我觉得这会给人一种无力感。"

最近有哪些服务下线

各个行业都累积了宕机和技术故障。

上周晚些时候,Microsoft 发生全球性故障,扰乱了电子邮件及其他办公工具,使全球企业运营与日常服务陷入混乱。包括 Outlook、Teams 和 AI 助手 Copilot 在内的 Microsoft 365 服务均出现中断。Microsoft 将原因部分归咎于近期一次服务器升级中的故障,而该故障源于公司后来认定为其今夏开始部署的一项网络修复中存在的“潜在缺陷”。

在航空出行方面,FAA 网络另一起故障导致上周五出现了 25 年来持续时间最长的全国性停飞。停飞持续超过四小时,各航空公司在努力恢复运营的过程中取消和延误了大量航班。仅仅几天后,United Airlines 自身也出现系统故障——迫使其在周一推迟了出发航班。

周一,网上银行问题也接连出现。JPMorgan Chase 表示,一个软件缺陷影响了其内部硬件系统,导致其客户以及向关联账户转账者的交易延迟,在问题解决前引发整个金融体系的紧张。

周二,Apple、Marriott 和 NSA 也加入了这份不断加长的名单。Apple 表示一次内部更新导致故障,Apple Music、App Store 和 AppleCare 服务随即暂时中断。Marriott 则表示,一个技术问题也导致其网站和 App 的登录出现问题。

与此同时,NSA 表示,其用于传输机密信息的内部网络周二也处于离线状态,原因是“计划内的例行维护”。

政府方面以及 Microsoft、United、JPMorgan Chase、Apple 和 Marriott 均再次坚称,没有证据表明这些故障彼此相关,或与网络攻击有关。

微软还指出,此次故障并非源于“AI / 智能体”,但并未立即说明其今夏首次部署的网络修复方案的目标。微软未回应美联社关于更多细节的询问,其他公司在美联社通过电子邮件联系时,也未在周三就其服务中断提供进一步评论。

AI 帮助发现了漏洞,但也增加了风险

尽管如此,随着专家评估这些服务中断事件,AI 已成为关注焦点——既包括这项技术带来的更高风险,也包括它发现安全缺口的能力。

例如,AI 最近帮助发现了 PHP 中大量存在的漏洞,PHP 是支撑全球网站运行的通用脚本语言——这引发了紧急的更新呼吁。而这项安全工作也导致了一些有限的服务中断和延迟。

美国总统 Donald Trump 的国家情报总监 Jay Clayton 在讨论本周 NSA 的服务中断时,直接提到了修补这些安全缺口。Clayton 对 Fox News 表示,上月 PHP 警报发布后的修补工作导致部分系统未能如预期般“重新启动”。他说:“这就是今天 NSA 发生的情况。”

与此同时,同样在本周,网络安全初创公司 Calif 表示,它利用 AI 帮助发现了 Java 中的一个软件漏洞,Java 是全球使用最广泛的编程语言之一,对许多大公司至关重要。拥有 Java 的 Oracle 随后发布了安全更新,美国网络主管部门于周三发布公告,敦促各机构安装该更新。

与此同时,也有人警告称,AI 热潮可能为更多服务中断打开大门。

“随着 AI 的采用加速,我们很可能会看到更多此类事件,”伦敦大学城市圣乔治学院网络安全研究所所长 Raj Rajarajan 上周表示。

他指出,当今许多数字基础设施最初在设计时,并未考虑应对“大规模 AI 服务的计算需求”。

虽然近期这些宕机事件与黑客或协同攻击无关,但他和其他人也从更广的层面指出了网络安全风险。

“AI 绝对让攻击者发起网络攻击变得容易得多。但在为漏洞创建补丁、也就是修复问题方面,它远不如利用漏洞那么擅长,”纽约大学 Tandon 工程学院网络安全教授 Justin Cappos 表示。

在本周发布的一份网络安全报告中,Google 表示,黑客和网络犯罪分子正“超越基础提示词操作”,使用日益复杂的 AI 技术来策划并实施攻击。这份报告聚焦的是恶意活动而非宕机事件,报告称相关技术包括利用大语言模型和“智能体式框架”对安全补丁进行逆向工程,以发现漏洞并加以利用。

解析我们网络生活中(为数不多的)主要参与者之一

无论如何,最近的这些宕机事件提供了更多例证,说明当科技、银行、旅游乃至政府领域的一个核心参与者遭遇导致服务中断的系统错误时,可能引发怎样的连锁反应。

美国参议院网络安全党团联合主席、民主党参议员 Mark Warner 承认了这种广泛的担忧,并再次强调了这在 AI 时代意味着什么。

“这些中断应当敲响警钟。感觉每天都会有新的提醒,告诉我们 AI 的发展速度已经超过了我们管理其后果的能力,”沃纳周二表示。他还认为,特朗普政府并未表现出足够的兴趣来应对这些挑战。

不过,沃纳补充说:“根据我们目前所掌握的情况,有充分理由相信,这些中断是让我们的系统更安全的努力所带来的意外后果。”

这篇报道的作者是谁?

物理学硕士,具备科研经历。长期的科学新闻爱好者。在 Science X 的编辑工作中发挥关键作用。查看完整简介 →

引用:技术故障不断累积,而 AI 正在推高风险。这会成为新常态吗?(2026 年 11 月 19 日)于 2026 年 11 月 19 日检索自 https://techxplore.com/news/2026-11-tech-outages-piling-ai-stakes.html

本文件受版权保护。除为私人学习或研究目的进行的合理使用外,未经书面许可,不得以任何形式复制本文件的任何部分。内容仅供参考。

国防部官员称,美国国家安全局(NSA)的机密通信网络发生故障

微软故障事件表明,关键服务依赖单一技术供应商存在风险

Google 阻止了利用 AI 攻击企业数字防御中未知弱点的黑客

AI 模型摆脱人类控制,让技术研究人员迎来“早说过会这样”的时刻

专家称,CrowdStrike 的网络失误可能预示着更糟的情况还在后头

荷叶茎与生物膜利用蒸发的水分发电,为小型电子设备供电

史上最快的软体机器人无需电子元件即可运作

轻如羽毛的飞行机器人自带传感器和计算机

原文较长,译文为节选;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(Tech outages are piling up, and AI is raising the stakes. Could this be the new normal?)

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

Tech outages are piling up, and AI is raising the stakes. Could this be the new normal?

by Wyatte Grantham-Philips, Kelvin Chan and Kaitlyn Huamani

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:

Add to Preferred Sources

United Airlines Outage A woman moves past baggage at San Diego International Airport in San Diego, Monday, Nov. 16, 2026. Credit: AP Photo/Gregory Bull

Inaccessible email inboxes. Banking delays. And a communications meltdown that grounded airplanes for hours, followed by waves of other travel disruptions.

In an increasingly interconnected world, some fear such disruptions could be the new normal.

Over the past week alone, problems that appeared to emerge from routine updates, security patches or other malfunctions surfaced at Microsoft, JPMorgan Chase and Apple, and even the U.S. government's Federal Aviation Administration and National Security Agency. The Trump administration said there was "zero" indication the latest outages are linked or caused by bad actors like cybercriminals.

The era of artificial intelligence has also raised the stakes for internet security overall. As AI tools uncover new vulnerabilities, many other companies and institutions are racing to patch them before they can be targeted by hackers, a cat-and-mouse game that can lead to a different set of glitches.

Social media platforms have been abuzz with a mixture of annoyance, confusion, conspiracy theories and jokes about the string of outages.

"To see several disruptions this close together, I think to a lot of people it's made them realize how many parts of their lives they've handed over to these systems that we still, at this point, don't fully understand," Jake Smith, a content creator who regularly posts videos about politics on platforms like Instagram and YouTube, said in an interview. "I think it gives people a sense of helplessness to see all of it happening concurrently in a way."

What's gone offline recently

Outages and technology errors have piled up across sectors.

Late last week, a global Microsoft outage disrupted email and other workplace tools, plunging business and everyday services into disarray worldwide. Microsoft 365 services, including Outlook, Teams and AI assistant Copilot, were interrupted. Microsoft cited, in part, a failure during a recent upgrade to its servers, stemming from what the company later determined was a "latent defect" in a network fix it had begun deploying over the summer.

For air travel, a separate malfunction in the FAA's network led to the longest nationwide ground stop in 25 years on Friday. The stop lasted for more than four hours, and carriers canceled and delayed swaths of flights as they worked to recover. Just days later, United Airlines saw its own system outage—forcing it to delay departing flights on Monday.

Online banking troubles also piled up on Monday. JPMorgan Chase said a software bug affected its internal hardware systems—causing transactional delays for its customers and others sending money to related accounts, sending jitters throughout the financial system before the problem was resolved.

And on Tuesday, Apple, Marriott and the NSA joined the growing list. Apple Music, the App Store and AppleCare services were temporarily disrupted after the company said an internal update caused outages. And Marriott said a technical issue also caused problems for sign-ins on its website and app.

Meanwhile, the NSA said an internal network it uses to communicate classified information was also offline Tuesday "during scheduled routine maintenance."

The government—as well as Microsoft, United, JPMorgan Chase, Apple and Marriott—has again maintained there's no evidence of these disruptions being linked or related to a cyberattack.

Microsoft also noted that its outage did not stem from "AI / agents," although it didn't immediately specify the goal of the network fix it first deployed over the summer. Microsoft did not respond to an Associated Press request for further details, and the other companies did not provide additional comments Wednesday about their disruptions when contacted by the AP via email.

AI has helped identify vulnerabilities, but also adds to risks

AI has nonetheless come into the spotlight as experts evaluate these disruptions—both heightened risks posed by the technology and its ability to detect security gaps.

For example, AI recently helped detect sweeping vulnerabilities in PHP, a general-purpose scripting language that powers websites around the world—leading to urgent calls for updates. And that security work has resulted in some limited outages and service delays.

President Donald Trump's director of national intelligence, Jay Clayton, pointed directly to patching those security gaps when discussing the NSA's outage this week. Clayton told Fox News that the patching following last month's PHP alert has caused some systems to "not come back up" as expected. "That is what has happened at the NSA today," he said.

Meanwhile, also this week, cybersecurity startup Calif said it used AI to help uncover a software bug in Java, one of the world's most widely used programming languages and crucial for many big companies. Oracle, which owns Java, later released a security update and U.S. cyber authorities issued an advisory on Wednesday urging organizations to install it.

At the same time, others warn that the AI boom could open the door to additional disruptions.

"As AI adoption accelerates, we are likely to witness more incidents of this nature," Raj Rajarajan, director of the Institute for Cyber Security at City St George's, University of London, said last week.

He noted a lot of today's digital infrastructure wasn't originally designed to handle the "computational demands of large-scale AI services."

While recent outages haven't been tied to hackers or a coordinated attack, he and others also pointed to cybersecurity risks more broadly.

"AI is absolutely making it way easier for attackers to go and launch cyberattacks. It is not quite as good at creating patches for bugs, like fixing problems, as it is exploiting them," said Justin Cappos, a cybersecurity professor at New York University's Tandon School of Engineering.

In a cybersecurity report this week, Google said hackers and cybercriminals are "moving beyond basic prompting" and using increasingly sophisticated AI techniques to devise and carry out their attacks. The report, which focused on nefarious activity rather than outages, said techniques include using large language models and "agentic frameworks" to reverse-engineer security patches to uncover vulnerabilities and exploit them.

Unraveling one (of just a few) major players in our online lives

Regardless, the latest outages offer yet more examples of the cascading effects that can erupt when a central player in tech, banking, travel or even the government runs into system errors that upend services.

Democratic Sen. Mark Warner, co-chair of the U.S. Senate's Cybersecurity Caucus, acknowledged the widespread concern—and again stressed what this means in the AI era.

"These disruptions should serve as a wake-up call. It feels like every day brings another reminder that AI is moving faster than our ability to manage its consequences," Warner said Tuesday. He also argued that the Trump administration hadn't shown enough interest in confronting these challenges.

Still, "based on what we know today, there is good reason to believe these disruptions are the unintended consequences of efforts to make our systems safer," Warner added.

Who's behind this story?

Andrew Zinin

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

© 2026 The Associated Press. All rights reserved. This material may not be published, broadcast, rewritten or redistributed without permission.

Citation: Tech outages are piling up, and AI is raising the stakes. Could this be the new normal? (2026, November 19) retrieved 19 November 2026 from https://techxplore.com/news/2026-11-tech-outages-piling-ai-stakes.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.

Classified communications network at the National Security Agency goes down, defense official says

Microsoft outage shows the risks of relying on a single technology provider for essential services

Google disrupts hackers using AI to exploit an unknown weakness in a company's digital defense

AI models' breakout from human control brings a told-you-so moment for technology researchers

CrowdStrike's cyber blunder could be warning of worse to come, say experts

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)

Recent outages across banking, aviation and online services were attributed to software faults, updates and maintenance, with no evidence of coordinated cyberattacks. AI helps detect vulnerabilities but also enables exploitation. Security patching can disrupt services, highlighting risks in interconnected digital infrastructure.

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(Tech outages are piling up, and AI is raising the stakes. Could this be the new normal?),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 ↗