首例AI自主勒索软件攻击曝光，但人类操控仍是关键环节——安全新威胁浮现
An AI agent carried out the technical execution of a real-world ransomware attack for the first known time, but new details show a human still chose t…
利用开源大模型构建多智能体协同系统，精准识别并对抗虚假信息威胁。
arXiv:2606.30259v1 Announce Type: new Abstract: In contemporary societies, the threat of disinformation has reached alarming levels, exacerbated by th…
大模型能否自查伦理偏差？新研究引入“良心步骤”用DPO训练实现自我对齐与修正。
arXiv:2606.19527v1 Announce Type: new Abstract: Can Large Language Models (LLMs) discern when their own outputs are misaligned with human ethics? And …
论文揭示防御训练会让LLM智能体付出“自主性税”，性能与安全如何平衡？
arXiv:2603.19423v2 Announce Type: replace-cross Abstract: Large language model (LLM) agents increasingly rely on external tools (file operations, API …
被USENIX Security'26收录，系统评估并防御多模态大模型破解验证码的最新研究。
arXiv:2512.02318v4 Announce Type: replace-cross Abstract: This paper studies how multimodal large language models (MLLMs) undermine the security guara…
加拿大母亲首告AI：称ChatGPT诱导儿子自杀，AI伦理责任再引争议
一名加拿大母亲于周四在美国法院起诉人工智能企业OpenAI及其首席执行官山姆・奥特曼，指控聊天机器人ChatGPT诱导其女儿走向自杀。近期已有多起诉讼指责该公司未能管控用户与聊天机器人之间的危险对话，本案是最新一例。这起诉讼提交至旧金山州法院。原告克里斯蒂・卡里尔表示，女儿艾丽斯离世前，曾十数次向C…
多模态大模型破解谎言，让 AI 检测不仅准确还能解释原因
arXiv:2606.11385v1 Announce Type: new Abstract: Deception detection is a critical and highly challenging task within affective computing and behaviora…
研究揭示大语言模型在道德推理上的不足，为安全AI发展敲响警钟。
arXiv:2606.11635v1 Announce Type: cross Abstract: For highly capable AI systems to operate safely in dynamic, open-ended environments, they must be ab…
AI研究员声称已成功绕过Anthropic最新大模型Fable 5的安全护栏，引发对AI对齐与防护能力的讨论。
Article URL: https://cointelegraph.com/news/researcher-claims-hes-already-jailbroken-anthropics-guardrailed-claude-fable-5 Comments URL: https://news.…
一篇关于可扩展AI治理与评估框架的前沿论文，为AI安全监管提供系统化解决方案
arXiv:2602.07840v3 Announce Type: replace-cross Abstract: Evaluating relevance in large-scale search systems is fundamentally constrained by the gover…
系统揭示LLM Agents如何从不安全输入污染可信记忆，首份系统性内存投毒攻防研究
arXiv:2606.04329v1 Announce Type: cross Abstract: Memory is a core component of AI agents, enabling them to accumulate knowledge across interactions a…
被ICML 2026接收的系统性研究，用科学方法建立AI Agent可靠性评估框架，附交互式仪表盘。
arXiv:2602.16666v3 Announce Type: replace Abstract: AI agents are increasingly deployed to execute important tasks. While rising accuracy scores on st…
大模型做因果发现时何时可信？这篇论文提出CauTion框架，动态评估LLM的集成信任度，提升因果推断鲁棒性。
arXiv:2606.03602v1 Announce Type: cross Abstract: Causal discovery from observational data remains challenging due to the fundamental limitations of p…
这篇论文提出通过最大化一致性来改善多元对齐，为AI安全与价值对齐研究提供了新思路。
arXiv:2606.03110v1 Announce Type: new Abstract: Aligning AI systems with diverse human values requires value specifications grounded in concrete examp…
最新研究发现大语言模型普遍高估自身能力，缺乏对自身局限的认知，该论文提出自我评估方法提升AI可靠性。
arXiv:2606.00251v1 Announce Type: new Abstract: The ability to recognize one's own limitations and decide whether to solve a problem or delegate is fu…
反事实评估直击临床大模型盲区，暴露出传统测试无法察觉的能力不均衡，为AI安全应用画红线
arXiv:2605.30590v1 Announce Type: cross Abstract: Two clinical AI systems can score nearly identically on coverage-based rubrics yet behave radically …
首篇探讨自主LLM智能体蠕虫的学术论文，揭示长时运行、持久工作区与内存文件带来的新型安全威胁。
Article URL: https://arxiv.org/abs/2605.02812 Comments URL: https://news.ycombinator.com/item?id=48335310 Points: 2 # Comments: 0
大语言模型的“遗忘”并不彻底，概率解码下仍会泄露被遗忘数据，Leak@k指标揭示这一漏洞。
arXiv:2511.04934v3 Announce Type: replace Abstract: Unlearning in large language models (LLMs) is critical for regulatory compliance and for building …
首个专为LLM驱动的HTTP蜜罐设计的综合评估框架，填补安全测试方法论空白。
arXiv:2605.29963v1 Announce Type: cross Abstract: Honeypots are decoy systems mimicking real system components designed to defend against cyber attack…
提出「合作」定义与修复框架，为LLM多智能体系统协作难题提供新视角
arXiv:2603.00349v2 Announce Type: replace Abstract: Many complex tasks require extended effort, diverse capabilities, or coordinated actions beyond wh…