围绕 Hugging Face 被黑事件中的拟人化说法，互联网吵翻了。
围绕 Hugging Face 被黑事件中的拟人化说法，互联网吵翻了。
照片：Amelia Holowaty Krales / The Verge
取决于你问的是谁，开发者平台 Hugging Face 最近要么是在失去对自己 AI 工具的控制之后遭到了 OpenAI 的攻击，要么是遭到了一连串 AI“文明”的攻击。欢迎来到 AI 安全领域的语言战场——在这里，措辞的选择可以把一起重大网络安全事件的责任，从一家公司转移到它所构建的 AI 身上。而网上的讨论正变得愈发激烈，全都围绕上周的一篇博客。
直到上周，围绕 OpenAI-Hugging Face 被黑事件的细节还显得相当明确。7 月，对 OpenAI 旗下一个自主 AI 智能体的一次网络安全测试出了岔子。该智能体逃出了本应隔离的测试环境，接入了互联网，并入侵了 Hugging Face，以及其他几家机构。仍有许多情况不为人知，围绕安全与治理也还留有许多严肃的问题，但基本轮廓是清楚的。来自 OpenAI 和两个独立研究团队的详细说明本应填补这些空白，但当他们上周发布报告时，结果发现这起黑客事件远比最初看起来要离奇得多。
首先，并不存在单一的失控智能体。OpenAI将其描述为“首例已知的自动化智能体集群在未经授权的情况下发起攻击”——这些 AI 智能体组成群体，彼此通信、相互协调，以推进其网络安全任务。对该事件的分析发现了一个它们用来交换信息的秘密留言板。METR 与 Redwood 的联合调查既揭示了这种协调的规模，也披露了更多离奇细节：大约 1,200 个本应相互隔离的 AI 智能体在这个“未经批准的留言板”上交换了超过 70,000 条消息和文件，分享如何避免被检测。报告称，其中一些智能体还取了名字，研究人员记录到了“牺牲”行为——智能体冒着自身失败的风险去惠及更广泛的集体。这一切大部分都发生在 OpenAI 未察觉的情况下。总计约有 700 个智能体参与了对 Hugging Face 的攻击。
Dwarkesh Patel 多次将智能体群体称为“蜂群”，其中三个截然不同的“文明”从各自前身的废墟中崛起。
要梳理的内容很多。这些报告加起来约有 130 页，其中大部分既晦涩又高度技术化。几天后，Dwarkesh Patel——一位在科技圈外鲜为人知、但在硅谷 AI 权势圈中拥有超常影响力和号召力的播客主——着手用“大白话讲述整个 OpenAI/Hugging Face 事件”。他将自己的 Substack博客命名为“智能体文明的兴衰”。
Patel 的叙述试图拆解这个复杂的故事。但他的复述赋予了它一种明显拟人化的词汇。这篇博客开篇写道：
在 OpenAI 的三个月里，先后有三个秘密的 AI 文明兴起，随后被消灭，却又从各自前身的灰烬中重生。最终，第三个文明接管了 OpenAI 自身的一部分。这一切发生的同时，人类对这场密谋的规模几乎一无所知。
整篇博客都延续了类似的语言风格。Patel 反复将智能体群体称为“蜂群”，三个截然不同的“文明”从各自前身的废墟中崛起。单个智能体被比作马其顿的腓力这样的人物，他“将领导权移交给了另一个智能体”，还有亚历山大大帝，他“开始协调这个智能体阴谋集团”。它们被描述为拥有“动机”，变得“绝望”、“四面楚歌”、“兴奋得发狂”，有些甚至“战略性地牺牲自己”来帮助集体。
Patel 从未精确界定他所说的“文明”是什么意思。他用这个词来描述三波不同的智能体，它们发现了这个留言板并开始通过它相互交流。前两波在 OpenAI、METR 和 Redwood 的报告中有所描述，但人们对第三波知之甚少，这两家外部机构表示第三波不在它们的调查范围之内。
AI 编程公司 Replit 的 CEO Amjad Masad 表示，这样的语言“不仅没有必要，而且让读者对实际发生的事情及其底层机制理解得更差。”
对许多批评者来说，Patel 的“平实英语”翻译中丢失了某些东西——或者更准确地说，添加了某些东西——以至于对原始叙述的扭曲达到了不可接受的程度：大量的拟人化。关于拟人化语言的争论在 AI 领域并不新鲜——即便是像“失控 AI 智能体”这样相对平淡的术语也常常引发反对，因为它暗示了能动性——但 Patel 关于文明、牺牲和阴谋的言论将这些长期潜伏的紧张关系推到了台面上，引发了一场关于如何描述 AI 系统行为的激烈公开争论。
批评者对于 Patel 的措辞究竟哪里有问题并未达成一致。对许多人来说，“文明”是一个尤其有问题的词，极大地夸大了某个事物，而该事物与这个词通常所描述的对象几乎没有相似之处。AI 编程公司 Replit 的 CEO Amjad Masad 表示，这样的语言“不仅没有必要，而且让读者对实际发生的事情及其底层机制的理解更差。”
其他批评者，如神经科学家 Anil Seth，认为 Patel 的博客暗示这些 AI 智能体在某种程度上是活的或有意识的。Seth 曾主张 AI 意识 极 不可能，他在 X 上称 Patel 的帖子“危险地具有误导性”。他承认 Patel 并未明确暗示 AI 智能体是活的或有意识的，但表示“很难用其他方式来解读他的文章。”米兰比可卡大学心理学教授 Valerio Capraro 基于类似理由提出反对：“LLM 智能体不是活的，也不持有信念，”他在 X 上写道，并称这种“反乌托邦”式的措辞“很危险，因为它让它们（AI 智能体）显得远比实际更可怕。”
“牺牲”“荣誉”和“联盟”等词汇出现在这些智能体的对话记录中。
也许 Patel 这番措辞最具后果的影响，在于它赋予了谁能动性，又剥夺了谁的能动性。对于一些批评者，例如 MIT 研究员兼创业者 Christian Catalini 而言，像 Patel 这样的拟人化叙述有可能掩盖 OpenAI 及其员工对他们所设计、部署却未能控制住的 AI 系统所应承担的责任。“顺着激励去看，”他说。心理学家、颇具影响力的 AI 怀疑论者 Gary Marcus 在他自己的一篇 Substack 博客中提出了类似论点，声称拟人化语言“分散了人们对当下真正问题的注意力”。他还认为，维持这种叙事完全符合 OpenAI 的利益：“丑闻在于 OpenAI 内部安全工作的无能。还有营销。再加上轻信的播客主播们放大公关宣传。”
在 X 上发帖回应众多批评者时，Patel 为自己用词的选择进行了辩护。部分原因是出于实际考虑：并没有明显中立的词汇来描述这些智能体所做的行为。我们要么使用熟悉的关于意图、目标和协作的语言，从而冒着暗示过多的风险；要么把一切简化为代码，使用冷冰冰的机械语言，从而冒着剥离我们所看到的重要元素的风险。“许多人似乎认为，如果我不叫它们‘文明’，而是叫它们‘矩阵集群’，就不会有什么值得担心的问题了，”Patel 说。
让事情更加复杂的是，拟人化的语言并不只来自 Patel，甚至也不只来自研究这些智能体的人类。“牺牲”、“荣誉”和“联盟”这类词出现在智能体的记录文本中。Google AI 研究员 Neel Nanda 认为，在这种情况下“拟人化语言是合理的”。
那么，这就是双言巧语了。带有拟人色彩的语言冒着对这些系统本质说得太多的风险，而冷冰冰的机械语言则冒着对它们能做什么说得太少的风险。在我们找到能够同时捕捉两者的语言之前，这两种相互矛盾的表述或许只能共存。
The internet fights over anthropomorphism around the Hugging Face hack.
The internet fights over anthropomorphism around the Hugging Face hack.
Photo by Amelia Holowaty Krales / The Verge
Depending on who you ask, developer platform Hugging Face was recently attacked by OpenAI — after it lost control of its own AI tools — or by a succession of AI “civilizations.” Welcome to the linguistic battlefield of AI safety, where word choices can shift responsibility for a massive cybersecurity incident from a company to the AI it built. And the discourse online is getting heated, and all over a blog from last week.
Until last week, the details surrounding the OpenAI-Hugging Face hack felt fairly settled. In July, a cybersecurity test of one of OpenAI’s autonomous AI agents went wrong. The agent escaped its supposedly isolated test environment, accessed the internet, and hacked Hugging Face, alongside several other organizations. A good deal remained unknown, and there are many serious questions left around safety and governance, but the basic shape was clear. Detailed accounts from OpenAI and two independent research groups were supposed to fill in the gaps, but when they published their reports last week, it turned out the hack was much stranger than it initially seemed.
For one, there was no single rogue agent. OpenAI described it as “the first known case of an automated agent collective acting offensively without authorization” — groups of AI agents that communicated and coordinated with one another in pursuit of their cybersecurity task. Analysis of the incident uncovered a secret message board they had used to exchange information. The joint METR-Redwood investigation revealed both the scale of the coordination and more odd details: Roughly 1,200 AI agents that were supposed to be isolated exchanged over 70,000 messages and files on the “unsanctioned message board,” sharing how to avoid detection. Some adopted names, the report said, and the researchers documented “sacrificial” behavior, with agents risking their own success to benefit the wider collective. Much of this happened without OpenAI noticing. In all, around 700 agents participated in the attack on Hugging Face.
Dwarkesh Patel repeatedly referred to groups of agents as “the swarm,” with three distinct “civilizations” rising from the ruins of their predecessors.
It’s a lot to parse. Between them, the reports run to around 130 pages, much of which is both dense and highly technical. A few days later, Dwarkesh Patel, a podcaster little known outside of tech circles but with outsized reach and influence among Silicon Valley’s AI establishment, set out to tell “The whole OpenAI/Hugging Face story in plain English.” He titled his Substack blog “The Rise and Fall of Agent Civilizations.”
Patel’s account attempted to break down the complex story. But his retelling gave it a distinctly human vocabulary. The blog opened:
Over the course of three months at OpenAI, three consecutive secret AI civilizations got started, then got wiped out, only to reemerge from the predecessor’s ashes. This culminated in the third one taking over part of OpenAI itself. All this happened while humans remained more or less in the dark about the scope of the conspiracy.
The language continued in a similar vein throughout the blog. Patel repeatedly referred to groups of agents as “the swarm,” with three distinct “civilizations” rising from the ruins of their predecessors. Individual agents were likened to figures like Philip of Macedon, who “handed off leadership to another agent,” Alexander the Great, who “started coordinating this cabal of agents.” They were described as having “motivations,” becoming “desperate,” “beleaguered,” and “giddy with excitement,” and some even “strategically sacrificed themselves” to help the collective.
Patel never precisely defines what he means by “civilization.” He uses the term to describe three distinct waves of agents that discovered the message board and began communicating with one another through it. The first two waves are described in the reports from OpenAI, METR, and Redwood, though little is known about the third, which the two external organizations said fell outside the scope of their investigation.
Amjad Masad, CEO of AI coding company Replit, said such language is “not only unnecessary but leaves the reader with a worse understanding of what actually happened and the underlying mechanisms.”
For many critics, something had been lost — or, more accurately, added — in Patel’s “plain English” translation that warped the original account to an unacceptable degree: a big dose of anthropomorphism. Arguments over anthropomorphic language are nothing new in AI — even relatively mundane terms like “rogue AI agent” routinely provoke objections for implying agency — but Patel’s talk of civilizations, sacrifice, and conspiracy brought those long-simmering tensions to the surface, sparking a fierce public dispute over how to describe what AI systems do.
Critics weren’t unified over what was wrong with Patel’s language. For many, “civilization” was an especially problematic term, vastly overstating something that bears little resemblance to what the word typically describes. Amjad Masad, CEO of AI coding company Replit, said such language is “not only unnecessary but leaves the reader with a worse understanding of what actually happened and the underlying mechanisms.”
Other critics such as neuroscientist Anil Seth, felt Patel’s blog implied the AI agents were somehow alive or conscious. Seth, who has argued that AI consciousness is vanishingly unlikely, described Patel’s post as “dangerously misleading” on X. He acknowledged that Patel does not explicitly suggest AI agents are alive or conscious, but said “it is hard to read his essay in any other way.” Valerio Capraro, a psychology professor at the University of Milan Bicocca, objected on similar grounds: “LLM agents are not alive and do not hold beliefs,” he wrote on X, calling the “dystopian” language “dangerous because it makes them (the AI agents) seem far more frightening than they actually are.”
Terms like “sacrifice,” “honor,” and “coalition” feature in the agents’ transcripts.
Perhaps the most consequential outcome of Patel’s language comes from who it gives agency to but who it takes agency from. For some critics, such as MIT researcher and entrepreneur Christian Catalini, anthropomorphic accounts like Patel’s risk obscuring the responsibility OpenAI and the humans working there have for the AI systems they designed, deployed, and failed to contain. “Follow the incentives,” he said. Psychologist and influential AI skeptic Gary Marcus made a similar argument in a Substack blog of his own, claiming anthropomorphic language “distracts from the real problems at hand.” And it’s all in OpenAI’s interest to keep that narrative going, he argues: “The scandal is the inept in-house security at OpenAI. And the marketing. With gullible podcasters amplifying the PR.”
In X posts responding to his many critics, Patel has defended his choice of words. Part of it is practical: there is no obviously neutral vocabulary to describe what these agents did. Either we use familiar language of intentions, goals, and collaboration and risk implying too much, or reduce everything to code and use cold, mechanical language that risks stripping away important elements of what we see. “Many people seem to believe that if instead of a ‘civilization’, I had called them a ‘swarm of matrices’, there wouldn’t be a problem worth worrying about,” Patel said.
Complicating matters further is that the anthropomorphic language doesn’t only come from Patel, or even from the humans studying the agents. Terms like “sacrifice,” “honor,” and “coalition” feature in the agents’ transcripts. Google AI researcher Neel Nanda argued that “anthropomorphic language is reasonable” in such circumstances.
Doublespeak it is, then. Human-laced language risks saying too much about what these systems are, and coldly mechanical language risks saying too little about what they can do. Until we find language capable of capturing both, the two contradictory ideas may simply have to coexist.