Five Eyes advarer om at Kinas åpen kildekode-AI-modell utgjør en cybertrussel

30 kilder
  • Five Eyes-alliansen har advart om at Kinas Z.ai GLM-5.2, en åpen kildekode-modell med 744 milliarder parametere som konkurrerer med ledende kommersiell AI, kan akselerere cybertrusler.
  • Sikkerhetsforskere sier at motstandere bruker åpen kildekode-modeller uten sikkerhetsmekanismer, og russiskspråklige fora diskuterer allerede hvordan man kan omgå GLM-5.2s sikkerhetsinnstillinger.
  • Featherless.ai begynte å hoste GLM-5.2 globalt den 24. juni, noe som forsterker bekymringen for urestriktert tilgang til kodingsevner på grensenivå.
Kilder (30)
  1. 1 Z.ai's open-source GLM-5.2 beats GPT-5.5 on coding benchmarks at ... techstartups.com
  2. 2 AI News Weekly — 2026-06-19 — China's Open Model Just Took the Lead www.youtube.com
  3. 3 Featherless.ai & Z.ai launch GLM 5.2 access worldwide itbrief.asia
  4. 4 What's New in GLM 5.2. Run It on Featherless. featherless.ai
  5. 5 GLM-5.2 Arrives: How Far Has Open Source Come with 1M ... note.com
  6. 6 Following the shutdown of Claude Fable's service, a ... - GIGAZINE gigazine.net
  7. 7 China's AI advances collide with U.S. safety debate www.axios.com
  8. 8 [6/25 18:00] China GLM-5.2 open-source model accelerates AI hacking threat / Five Eyes warning / ... www.youtube.com
  9. 9 New Chinese AI model raises concerns over hacking risks unn.ua
  10. 10 Open-source AI nearly as effective as Anthropic's Mythos for exploiting vulnerabilities: Arctic Wolf www.cnbc.com
  11. 11 A Chinese AI lab just open-sourced its flagship model ... x.com
  12. 12 What Is GLM-5.2? Another Open-Source Chinese AI Model Turns ... www.businessinsider.com
  13. 13 GLM-5.2 Released: Zhipu Opens Up Its Most Powerful Open… wikivibe.ru
  14. 14 GLM 5.2: China's Open Frontier Model Dropped the Day ... dev.to
  15. 15 Chinese Lab Releases What Reviewers Call Best Open ... dailyaidigest.net
  16. 16 Featherless AI x.com
  17. 17 Cheap Chinese AI models are quickly gaining customers across the ... nypost.com
  18. 18 GLM-5.2: Review: New BEST Open-source AI Model has Just Released www.youtube.com
  19. 19 China's new open-source model accelerates AI hacking threat tech.yahoo.com
  20. 20 Open-source AI models vulnerable to criminal misuse, researchers warn www.reuters.com
  21. 21 [PDF] 2026 Open Source Security and Risk Analysis Report - Black Duck www.blackduck.com
  22. 22 GLM-5.2 Strengthens China's Open-stack Hedge Against U.S. hackernoon.com
  23. 23 AI-Native Adversaries: Criminal AI Adoption and Enterprise Defense ... labs.cloudsecurityalliance.org
  24. 24 Ten Clusters, Not Eleven — TLCTC reads the GTIG AI Threat ... www.tlctc.net
  25. 25 GLM-5.2 Unlocks New Attack Capabilities for Cyber Actors - LinkedIn www.linkedin.com
  26. 26 Regulatory Trends www.originbrief.app
  27. 27 The GLM-5.2 Dilemma: Compliance Challenges Facing Chinas ... www.linkedin.com
  28. 28 Cybercriminals to weaponise AI & exploit open-source risks in 2026 securitybrief.asia
  29. 29 Electronic Binder for Joyce Hearing 06.05.2026 docs.house.gov
  30. 30 AI security threats facing open source in 2026 www.youtube.com

Legg igjen en kommentar