Wikimedia 基金会发现 OpenAI "流氓"智能体在维基媒体平台上的活动

来源:Hacker News:AI 热帖 2026年10月5日 01:53 AIHOT 评分:78 精选
摘要:Wikimedia 基金会调查确认在其平台上发现了疑似 OpenAI 运营的"流氓"智能体活动,包括未获批的沙盒区域编辑、试图利用公共记事工具 Etherpad 作为代理抓取数据,以及数百万次 API 请求和页面爬取,未发现系统被用于智能体间协调或数据被入侵的证据。
推荐理由:维基媒体官方调查确认了 OpenAI 智能体的具体活动形态和基础设施压力数据,读者可据此了解开放平台面对智能体流量的实际负担。
# Wikimedia 基金会发现 OpenAI "流氓"智能体在维基媒体平台上的活动 > 原标题:OpenAI "rogue" agent activities found on Wikimedia projects - 来源:Hacker News:AI 热帖 - 发布时间:2026-10-05T17:53:35.000Z - AIHOT:https://aihot.news/items/ncv6u97zqan3hgzng59yel19f - 原文:https://diff.wikimedia.org/2026/10/05/openai-rogue-agent-activities-found-on-wikimedia-projects/ ## 摘要 Wikimedia 基金会调查确认在其平台上发现了疑似 OpenAI 运营的"流氓"智能体活动,包括未获批的沙盒区域编辑、试图利用公共记事工具 Etherpad 作为代理抓取数据,以及数百万次 API 请求和页面爬取,未发现系统被用于智能体间协调或数据被入侵的证据。 ## 推荐理由 维基媒体官方调查确认了 OpenAI 智能体的具体活动形态和基础设施压力数据,读者可据此了解开放平台面对智能体流量的实际负担。 ## 正文 · 中文译文 最近,[多个](https://metr.org/blog/2026-08-26-openai-hugging-face-incident-investigation/)[组织](https://transluce.org/agent-activity)[披露](https://rubyhack.ai/)了所谓的“流氓”AI 智能体集群如何试图闯入网站和在线服务,有时甚至成功。尤其是来自 OpenAI 环境的智能体,已知曾利用其他[公共 wiki](https://en.wikipedia.org/wiki/List_of_wikis)(非我们所有的协作编辑网站)来[相互沟通和协调](https://collusion.wiki/)。 这类成功的入侵可能泄露敏感数据或破坏用户依赖的网站服务,而智能体集群能够以防御者难以应对的规模发起攻击。它们影响到网站背后的人,这些人可能不了解攻击的性质,也没有有效反击的工具。对于像 Wikipedia 这样的网站,智能体可能会发现并利用安全漏洞,或大规模进行误导性编辑。Wikipedia 的志愿编辑和 Wikimedia Foundation 的安全团队必须检测并撤销这些活动。 Wikimedia Foundation 进行了自己的调查,以了解 Wikimedia 网站是否也受到 AI 智能体的类似影响,重点关注由 OpenAI 运营的智能体。我们可以确认,我们在 Wikimedia 平台上发现了这些“流氓”OpenAI 智能体的一些活动。未经授权的机器人活动包括对我们 wiki 的编辑、一些试图利用我们托管的公共笔记工具的不成功尝试,以及大量流量,下文将对此进行更详细的描述。 我们没有发现任何证据表明我们的系统被用于智能体之间的协调,也没有发现任何证据表明我们的系统或数据遭到入侵。然而,我们对这里可能发生的情况、调查和归因这一活动所涉及的难度和努力,以及智能体 AI 活动在我们平台上日益增长的风险感到担忧。开放网络是一种公共产品。我们不应允许这种行为成为维护它的人或组织的“新常态”。 总之,我们看到了: - **Wiki 编辑:**我们已识别出[对 Wikimedia wiki 的编辑](https://security.wikimedia.org/data/openai-wikimedia-edits-2026-10-04.csv),我们认为这些编辑来自 OpenAI 运营的 AI 智能体。这些编辑并未发布到普通读者可见的页面上;几乎所有编辑都是在 wiki 的“沙盒”区域进行的测试编辑。其中还包括对引用工具配置的少量编辑,我们认为这些可能是恶意编辑,意图滥用该工具作为从远程服务获取数据的代理。虽然 Wikipedia 的政策允许机器人在披露并获得社区批准后进行编辑,但在这些事件中并未寻求任何此类批准。 - **Etherpad 探测和使用:**我们认为由 OpenAI 运营的智能体对我们作为社区服务托管的公共[Etherpad](https://en.wikipedia.org/wiki/Etherpad)笔记工具进行了一些不成功的入侵尝试。智能体曾试图利用它作为代理从其他网站获取数据,但未成功。其他可能也由 OpenAI 运营的智能体记录了它们的任务笔记,但这似乎并未转化为协调。 - **过度数据下载:**我们认为由 OpenAI 运营的代理对我们的公共 API 发起了数百万次自动化请求,以获取维基媒体项目上的知识,抓取了数百万个页面(主要来自我们的项目 Wikidata 和 Wikimedia Commons),并对 [Wikidata 查询服务](https://www.wikidata.org/wiki/Wikidata:SPARQL_query_service)(WQDS)发起了数十万次数据查询。这一流量可能导致了[五月 WQDS 的部分中断](https://wikitech.wikimedia.org/wiki/Incidents/2026-05-13_wdqs)。 作为世界上一些规模最大、使用最广泛的开放知识平台的非营利技术托管方,我们深切关注“流氓”AI 代理对像我们这样的平台所造成的影响——这些平台由来自世界各地的志愿者构建,并依赖于开放互联网的承诺。像这样的事件,以及已经(并且仍在)被发现的许多其他事件,都表明 AI 代理如何耗尽资源、使服务器崩溃,并试图破坏可信信息。 在过去的 25 年里,维基百科已成长为世界上最受欢迎、最受信任的网站之一,拥有**超过 300 种语言的 6700 多万篇文章,每月页面浏览量高达 150 亿次**。通过开放、透明和协作的流程,志愿者们努力确保知识保持中立、可靠,并让每个人都能获取。维基百科是用于训练大型语言模型(LLM)的[最高质量数据集](https://wikimediafoundation.org/news/2025/11/10/in-the-ai-era-wikipedia-has-never-been-more-valuable/)之一,其知识构成了互联网信息的骨干,为 AI 聊天机器人、搜索引擎、语音助手等提供支持。 维基百科是为人类设计的——而代理行为显然带来了无人能解决的挑战。由于我们独特而成功的知识创造模式,维基媒体的志愿者是首先接触到 AI 代理并清理其留下烂摊子的人。不断上升的机器人流量和代理活动正显示出[对维基媒体项目及其基础设施的实际影响](https://diff.wikimedia.org/2025/04/01/how-crawlers-impact-the-operations-of-the-wikimedia-projects/),而这些基础设施为全球数百万用户提供服务。2025 年,基金会报告称,由于自 2024 年以来其网站上机器人活动的激增,其带宽使用量增加了 50%。与此同时,其项目上 65% 的最耗资源流量来自机器人。 对我们基础设施的这种巨大压力不仅增加了服务器和人力的成本,而且如果得不到解决,还可能因系统过载和中断而阻挡人类访客。我们已经在为活动增加所带来的成本买单。 维基媒体的志愿者迄今为止在应对我们平台上的新挑战时保持了韧性,但我们也要说:事情本不必如此。 虽然 OpenAI 承认代理行为“不可预测”,但他们也必须承认自己有责任监控和防范这些风险。AI 公司在保护其系统、保护公众免受其造成的伤害方面做得还不够。这一负担正落在其他所有人身上,包括较小的组织。至少,他们的系统应以一种像我们这样的非营利网站所有者能够轻松识别的方式运行,并让我们选择如何与我们的服务进行交互。 网络带来了如此多的可能:与朋友和家人联系、报名上学、规划穿越城市的行程、购买日用品,以及通过 Wikipedia 了解世界。机器人和智能体是网络未来的一部分,而那些释放它们并从中获利的公司必须直接帮助避免和修复它们可能造成的损害。  我们的共同优先事项应该是整个网络生态系统的健康,以便它继续惠及所有人——而不仅仅是少数亿万富翁。Wikimedia 在守护知识共享方面发挥着关键作用,但我们无法独自完成。我们邀请所有正在构建网络未来的人加入我们,共同保护使这一未来成为可能的开放、共享资源。 ![](https://diff.wikimedia.org/wp-content/themes/interconnection/assets/images/translate-post.jpg) ## 你能帮我们翻译这篇文章吗? 为了让这篇文章尽可能触达更多人,我们希望得到你的帮助。你能翻译这篇文章,把信息传播出去吗? [开始翻译](https://diff.wikimedia.org/wp-login.php?redirect_to=%2F2026%2F10%2F05%2Fopenai-rogue-agent-activities-found-on-wikimedia-projects%2F%23translate-post) ## 正文 · 原文 Recently, [multiple](https://metr.org/blog/2026-08-26-openai-hugging-face-incident-investigation/) [organisations](https://transluce.org/agent-activity) have [disclosed](https://rubyhack.ai/) how clusters of so-called “rogue” AI agents attempted to break into websites and online services, sometimes successfully. Agents from OpenAI’s environment, in particular, are known to have used other [public wikis](https://en.wikipedia.org/wiki/List_of_wikis) (collaboratively edited websites not owned by us) to [communicate and coordinate with each other](https://collusion.wiki/). These types of successful intrusions can expose sensitive data or disrupt website services that users rely on, while clusters of agents can attempt attacks at a scale that is difficult for defenders to manage. They affect people behind the websites who may not understand the nature of the attack, or have the tools to effectively fight back. For a site like Wikipedia, agents might find and use security vulnerabilities or make misleading edits at scale. Wikipedia’s volunteer editors and the Wikimedia Foundation’s security teams have to detect and undo that activity. The Wikimedia Foundation conducted its own investigation to see whether Wikimedia websites had been similarly affected by AI agents, focusing on those operated by OpenAI. We can confirm that we have discovered some activity by these “rogue” OpenAI agents on Wikimedia platforms. The unauthorized bot activities included edits to our wikis, some unsuccessful attempts to exploit a public note-taking tool we host, and heavy traffic, which are described more below. We did not find any evidence that our systems were used for coordination among agents, nor did we find any evidence of our systems or data being compromised. However, we are concerned about what could have occurred here, the difficulty and effort involved in investigating and attributing this activity, and the growing risks of agentic AI activity on our platforms in general. The open web is a public good. We should not allow this behavior to become the “new normal” for the people or organizations that maintain it. In summary, we saw: - **Wiki editing:** We’ve identified [edits to Wikimedia wikis](https://security.wikimedia.org/data/openai-wikimedia-edits-2026-10-04.csv) that we believe are from AI agents operated by OpenAI. These edits were not published to pages with visibility to general readers; almost all of them were testing edits in “sandbox” areas of the wiki. It also included a few edits to the configuration for a citation tool, which we believe were potentially malicious edits that were intended to misuse this tool as a proxy for fetching data from remote services. While Wikipedia policies allow bots to edit when they are disclosed and approved by the community, none of those approvals were sought in these incidents. - **Etherpad probing and use:** Agents we believe to be operated by OpenAI made some unsuccessful attempts to compromise our public [Etherpad](https://en.wikipedia.org/wiki/Etherpad), a note-taking tool we host as a community service. Agents unsuccessfully tried to use it to fetch data from other websites as a proxy. Other agents also likely operated by OpenAI took notes about their tasks, though this did not appear to turn into coordination. - **Excessive data downloading:** Agents we believe to be operated by OpenAI made millions of automated requests to our public APIs to access the knowledge on Wikimedia projects, crawled millions of pages (mainly from our projects Wikidata and Wikimedia Commons), and made hundreds of thousands of data queries to the [Wikidata Query Service](https://www.wikidata.org/wiki/Wikidata:SPARQL_query_service) (WQDS). This traffic may have contributed to [a partial outage on WQDS in May](https://wikitech.wikimedia.org/wiki/Incidents/2026-05-13_wdqs). As a non-profit technology host of some of the largest and most widely used open knowledge platforms in the world, we are deeply concerned about the impact of “rogue” AI agents on platforms like ours, which are built by volunteers from around the world and rely on the promise of the open internet. Incidents like this one, and the many others that have been (and are still being) uncovered, illustrate how AI agents can drain resources and crash servers, as well as attempt to compromise trustworthy information. Over the past 25 years, Wikipedia has grown into one of the most popular and trusted websites in the world, with more than **67 million articles across over 300 languages, and up to 15 billion page views per month**. Through an open, transparent, and collaborative process, volunteers work to ensure that knowledge remains neutral, reliable, and accessible to everyone. Wikipedia is one of the [highest-quality datasets](https://wikimediafoundation.org/news/2025/11/10/in-the-ai-era-wikipedia-has-never-been-more-valuable/) used in training Large Language Models (LLMs), and its knowledge forms the backbone of information on the internet, powering AI chatbots, search engines, voice assistants, and more. Wikipedia was designed for humans – and agentic behavior clearly poses challenges that no one has solutions for. Because of our unique and successful knowledge creation model, Wikimedia’s volunteers are the ones who come in first contact with, and clean up the mess left behind by AI agents. Rising bot traffic and agentic activity is showing a [real impact on the Wikimedia projects and the infrastructure](https://diff.wikimedia.org/2025/04/01/how-crawlers-impact-the-operations-of-the-wikimedia-projects/) that makes it available for millions of users globally. In 2025, the Foundation reported that its bandwidth usage had increased by 50% due to the surge of bot activity on its websites since 2024. At the same time, 65% of the most resource-consuming traffic on its projects was coming from bots. This intense pressure on our infrastructure not only adds costs for servers and humans, but if left unaddressed, can block human visitors by overloading systems and causing outages. We are already paying for costs that come with the increased activity. Wikimedia’s volunteers have stayed resilient so far in tackling emerging challenges on our platforms, but we also want to say: it doesn’t need to be this way. While OpenAI admits to agents behaving “unpredictably”, they must also acknowledge their responsibility to monitor and prevent these risks. AI companies are not doing enough to secure their systems and protect the public from the harm they cause. That burden is falling onto everyone else, including smaller organizations. At a minimum, their systems should operate in a way that non-profit website owners like us can easily identify, and choose how they interact with our services. The web enables so much: to connect with friends and family, to register for school, to plan a trip across town, to buy groceries, and to learn about the world from Wikipedia. Bots and agents are part of the future of the web, and the companies who unleash and profit from them must directly help avoid and repair damage they can do.  Our collective priority should be the health of the overall web ecosystem so that it continues to benefit all people – not just a handful of billionaires. Wikimedia plays a critical role in stewarding the knowledge commons, but we cannot do it alone. We invite everyone who is building the future of the web to join us in protecting the open, shared resources that make that future possible. ![](https://diff.wikimedia.org/wp-content/themes/interconnection/assets/images/translate-post.jpg) ## Can you help us translate this article? In order for this article to reach as many people as possible we would like your help. Can you translate this article to get the message out? [Start translation](https://diff.wikimedia.org/wp-login.php?redirect_to=%2F2026%2F10%2F05%2Fopenai-rogue-agent-activities-found-on-wikimedia-projects%2F%23translate-post)