这篇论文戳破了「模型安全=一切安全」的幻觉，做自主智能体开发、RAG 系统或浏览器自动化工具的团队，建议认真看看攻击面到底在哪。
Google DeepMind 最新论文首次系统分类了 6 种针对自主 AI 智能体的攻击类型，指出恶意网站可以检测到 AI 智能体并展示人类看不到的隐藏内容。这些攻击包括在 HTML 注释或白底白字文本中隐藏指令、图像像素隐写术、PDF 或元数据中的覆盖命令、跨会话持久化的记忆投毒、目标劫持以及多智能体设置中的级联攻击。论文强调，AI 智能体的真正安全问题不仅在于模型本身，更在于它所读取的环境——网络本身可以被武器化。在基准测试中，隐藏的提示注入在多达 86% 的场景中部分控制了智能体，子智能体劫持成功率 58-90%，数据外泄攻击在五种不同智能体架构中成功率超过 80%。
This Google DeepMind’s paper is a serious warning …
This Google DeepMind’s paper is a serious warning for anyone using autonomous agents today.
Gives the first clear taxonomy of 6 attack types where harmful websites can detect AI agents and show them hidden content humans never see, like
- Instructions buried in HTML comments or white-on-white text
- Steganography in image pixels
- Override commands in PDFs, metadata, or even speaker notes
- Memory poisoning that persists across sessions
- Goal hijacking and cross-agent cascades in multi-agent setups
The real security problem for AI agents is not just the model, but the environment it reads.
The web itself can be weaponized against autonomous AI agents. As agents increasingly browse the internet, read emails, execute transactions, and spawn sub-agents, the information environment becomes an attack surface.
In one cited benchmark, hidden prompt injections embedded in web content partially commandeered agents in up to 86% of scenarios, sub-agent hijacking working 58–90% of the time, and data exfiltration attacks clearing 80% across five different agent architectures.
That reframes the whole debate.
We usually talk about model safety as if the danger sits inside the weights, but agents do something more fragile: they browse, retrieve, remember, and act on untrusted material in real time.
Here’s the thing to worry about.
A web page does not have to look malicious to be dangerous to an agent, because the agent may parse what humans never see: hidden HTML comments, metadata, CSS-hidden text, formatting syntax, or adversarial content embedded in images and other media.
The threat gets more serious once memory enters the loop.
If an agent uses RAG or persistent memory, poisoning no longer has to win in one shot. It can sit quietly in a corpus or memory store and activate later, which is why the paper highlights results showing latent memory poisoning above 80% attack success with less than 0.1% data contamination.
---
ssrn .com/sol3/papers.cfm?abstract_id=6372438