Skip to content

Scan record

Everything the site collected, newest first by collection time (Beijing time), back seven days.

Nov 1640 loaded

  1. Amazon discloses AI agent attacks targeting applications such as LiteLLM and Dify, with API keys as the objective

    NewsThe Hacker NewsEventThe item you came from

  2. EDPB sets automated decision-making in recruitment (including AI candidate screening) as its 2027 coordinated enforcement topic

    NewsCASRAIGovernance

  3. Uber data breach: hackers breached AI vendor Fireworks to steal driver's license information

    NewsTechTimesEvent

  4. Malaysia identifies five categories of AI threats, including deepfakes and cyberattacks

    NewsNew Straits TimesGovernance

  5. After the affair scandal, Silicon Valley pushes for the return of former Nvidia CEO Jensen Huang; as 160,000 people petition, the US Congress expands its investigation

    NewsThe Asia Business DailyAI updates

  6. Hinton pushes back against Fei-Fei Li: those who say AI has "zero risk" are often the ones with vested interests in AI companies

    News澎湃新闻 · 检索「AI风险」Opinion

  7. METR research note: measuring identifiable giveaways of alignment evaluations

    NewsMETRAlignment

  8. METR measures the giveaways of alignment evaluations; investigation agents are increasingly able to distinguish alignment from capability evaluations

    NewsMETRAlignment

  9. AI wearable devices can evade radio frequency detectors, South Korea considers multi-layer exam prevention and control measures

    NewsChosunBizEvent

  10. AI exam cheating prompts South Korea to push for heavier penalties for the university entrance exam

    NewsChosunBizEvent

  11. South Korean police seized a miniature camera in the tax accountant examination cheating investigation

    NewsYonhapEvent

  12. Taiwan's Ministry of Justice Investigation Bureau busted the BossAI fire system hacking case; 3 suspects stole OHCA information and resold it to the funeral industry

    NewsiThome (Taiwan)Event

  13. Taiwan's Ministry of Digital Affairs: AI agent cyberattacks cause more than ten companies and platforms to report being affected, with cumulative affected personal data exceeding 1 million records

    NewsiThome (Taiwan)Event

  14. UK House of Lords report: government complacent, UK exposed to threat of multiple crises

    NewsThe Guardian · AIGovernance

  15. Personalizing the basic and moral values of an LLM-controlled assistant robot

    PaperarXiv: 其他 AI 安全论文

  16. Alignment through socialization: trade-off learning under selection pressure

    PaperarXiv: 其他 AI 安全论文

  17. Can mechanistic interpretability help explain LLM agent failures? A case study of web element hallucination

    PaperarXiv: 其他 AI 安全论文

  18. TradeGauntlet: Evaluating compliance controls of autonomous trading agents

    PaperarXiv: 其他 AI 安全论文

  19. Between obligations and evidence: a participatory study translating the EU AI Act into technical requirements

    PaperarXiv: 其他 AI 安全论文

  20. Causal AI-driven secure millimeter-wave beam alignment based on a defensive variational autoencoder

    PaperarXiv: 其他 AI 安全论文

  21. The Right Tokens, the Right Time: From Per-Token Gating to Trajectory-Aware Activation Steering

    PaperarXiv: 其他 AI 安全论文

  22. From Bytes to Behavior: Request-Level Attribution from Hidden States of Running GPUs

    PaperarXiv: 其他 AI 安全论文

  23. LLM puts the existential risk of AI at 10%

    PaperarXiv: 其他 AI 安全论文

  24. Forget the lineage: forward-secure, lineage-complete cryptographic erasure for LLM agent memory

    PaperarXiv: 其他 AI 安全论文

  25. Reasoning Models Confabulate Explanations to Defend Injected Beliefs About Their Own Requests

    PaperarXiv: 其他 AI 安全论文

  26. Adversarial attacks on symbolic vocabulary vulnerabilities in LLM planners

    PaperarXiv: 其他 AI 安全论文

  27. Cracks in the Embedding: Robustness and Stability of JEPA-Based Chest X-ray Representations

    PaperarXiv: 其他 AI 安全论文

  28. Soteria: Formally Verified Planning and Runtime Enforcement for Safe LLM Agents

    PaperarXiv: 其他 AI 安全论文

  29. Anti-distillation is cheap: lightweight distillation resistance via style disruption

    PaperarXiv: 其他 AI 安全论文

  30. Auditing Chatbots' Encroachment into the Professional Legal Practice Domain

    PaperarXiv: 其他 AI 安全论文

  31. Safety side effects of preventive personality steering: an empirical audit

    PaperarXiv: 其他 AI 安全论文

  32. System security is only as strong as the weakest agent: an attack that occurs only in multi-agent systems

    PaperarXiv: 其他 AI 安全论文

  33. Pushing Down the Mountain: The Effect of Framing Effects and Internal State Interventions on Collusive Behavior of Long-Horizon LLM Agents

    PaperarXiv: 其他 AI 安全论文

  34. Adversarial tracking: attacks on image sensors for autonomous vehicle object tracking

    PaperarXiv: 其他 AI 安全论文

  35. Unfinished-Bench: Measuring Rule-Violating Behavior of Persistent Agents in Real Work Environments

    PaperarXiv: 其他 AI 安全论文

  36. Physical Attacks on End-to-End Autoencoder-Based Communications via Evolutionary Transfer Learning

    PaperarXiv: 其他 AI 安全论文

  37. Google Project Zero will shorten vulnerability disclosure deadlines in stages: 60 days starting in 2027, 30 days starting in 2028

    NewsiThome (Taiwan)Governance

  38. New York Times: Lutnick family business's ties to Malaysian data centers raise conflict-of-interest questions

    NewsSeoul Economic DailyGovernance

  39. Cheating in China's college entrance exam can carry a maximum sentence of 7 years in prison; AI question-solving features suspended during exams

    NewsSeoul Economic DailyEvent

  40. Top scorer in South Korea's tax accountant exam admits to cheating with a miniature camera and generative AI and turns himself in

    NewsFinancial News KoreaEvent