Anthropic segist hafa stöðvað misnotkun gervigreindar í lífvopnarannsóknum

18 heimildir
  • Anthropic gaf út skýrslu um ógnargreiningu þar sem greint er frá misnotkun sem fyrirtækið stöðvaði á tímabilinu desember 2025 til ágúst 2026 í sjö flokkum skaðlegrar hegðunar.
  • Meðal mála voru tilraunir til að nota Claude við rannsóknir á virkniaukningu (e. gain-of-function) chikungunya-veirunnar, aðlögun fuglaflensu og styrkumsóknir vegna rannsókna á lifandi orthopox-veirum, að sögn fyrirtækisins.
  • Í skýrslunni er einnig fjallað um ríkistengda gerendur sem sjálfvirknivæddu netárásir, en hún kemur út örfáum dögum eftir að rannsakandi hjá Anthropic sagði upp störfum vegna áhyggna af öryggi gervigreindar.
Heimildir (18)
  1. 1 Countering misuse of AI: September 2026 / Anthropic www.anthropic.com
  2. 2 Is AI being used to develop more deadly chikungunya virus, bioweapons? What Anthropic’s latest report reveals www.theweek.in
  3. 3 Anthropic blocks possible attempt to use AI to make ... www.bbc.com
  4. 4 Anthropic says it blocked misuse of its AI that could have supported biological weapons techxplore.com
  5. 5 Anthropic exposes threats from bad actors using AI for bioweapons research, kamikaze drones (ANTHRO:Private) seekingalpha.com
  6. 6 Anthropic Documents Nine Months of AI Misuse Across ... aigovernance.com
  7. 7 Anthropic Says Claude AI Blocked Biological Weapons Attempts www.gadgetreview.com
  8. 8 September 2026 / Anthropic | Riggs Goodman III www.linkedin.com
  9. 9 Anthropic Threat Report: Claude Misuse Cases (Sept 2026 ... www.explainx.ai
  10. 10 Anthropic says it blocked misuse of its AI that could have ... ca.finance.yahoo.com
  11. 11 Anthropic says it blocked possible attempts to use AI to ... www.cnn.com
  12. 12 Anthropic threat report — attackers now let… - AI/TLDR ai-tldr.dev
  13. 13 Bioweapons research stopped by Anthropic, new report says www.usatoday.com
  14. 14 Anthropic's September 2026 Threat Intelligence Report: AI ... undercodetesting.com
  15. 15 Detecting and countering misuse of AI: August 2025 www.anthropic.com
  16. 16 Disrupting the first reported AI-orchestrated cyber ... www-cdn.anthropic.com
  17. 17 AI Gone Rogue - What Anthropic's Report Means for ... ironscales.com
  18. 18 AI Watch: Anthropic flags rising AI cyber threats; OpenAI launches ChatGPT for financial services www.cnbctv18.com