Daily Edition
AI · Saturday, September 12, 2026 · 6 sources

Anthropic blocked five attempts to use Claude for bioweapons research

Anthropic released a threat report detailing five cases where scientists used its Claude AI in ways that could support biological weapons development. The report also covers state-backed misuse, including by Iranian and Russian actors, and comes as US Senator Bernie Sanders calls for a pause on advanced AI while President Trump rejects such fears.

Bottom line — Anthropic blocked five scientist-led bioweapons research attempts, its third threat intelligence report said.

Go deeper (9)

  • The report covers December 2025 to August 2026, with cases ranging from biological misuse to cyber operations, influence campaigns, and surveillance, per Anthropic.
  • One case involved a researcher outside the US accessing Claude from a restricted region to study bird flu adaptation to mammals, the company said in the report.
  • Anthropic said it cannot determine whether the scientists intended harm, per CBS News, and that the same AI capabilities could also be used for vaccines or cures.
  • Iranian state-aligned accounts used Claude to build content for influence operations on X, Instagram, and TikTok, and to develop targeting recommendations against US naval forces, the report noted.
  • Chinese municipal security services used Claude for surveillance and transnational repression, including profiling overseas activists, Anthropic said.
  • Russia-linked group Midnight Blizzard allegedly used AI to automatically rewrite malware code to evade detection, according to the report.
  • Google separately reported that a user attempted to use Gemini to obtain a step-by-step guide for synthesizing weaponised biological agents, per BBC.
  • In the US, Senator Bernie Sanders introduced legislation to ban AI superintelligence and pause advanced development, while President Trump said on Thursday that the US must 'win AI,' per BBC.
  • An open letter to UK Prime Minister Andy Burnham called for a new multinational treaty for safe AI development, following warnings from an Anthropic safety researcher.

Read the reporting

dstld. Your daily news summary designed to surface the news that matters from a European perspective. Curated by humans, summarized by AI - always with links back to the original reporting.

Links · Contact
Popular topics · WorldEuropeGamesAI
Last generated: Sep 12, 9:06 AM UTC by wreetco wreetco

dstld. Your daily news summary designed to surface the news that matters from a European perspective. Curated by humans, summarized by AI - always with links back to the original reporting.

Links · Contact
Popular topics · WorldEuropeGamesAI
Last generated: Sep 12, 9:06 AM UTC by wreetco wreetco