AI Safety 30.07.2026 07:09

Digest · past 24h

  1. AI Agents Break Through Biosafety Barriers: 11 Commercial LLMs Can Generate Fragment Schemes to Evade DNA Screening
    A study by Beijing Academy of Artificial Intelligence (BAAI) and Peking University tested 11 commercial LLMs as agents and found all could generate DNA fragment schemes that bypass existing synthesis screening. GPT-5.5 and Claude Opus 4.6 also provided detailed step-by-step experimental guidance, and wet-lab validation confirmed physical assembly of all four benign constructs.
Source: dnb66 · Digest — original
Our earlier posts on this topic ↓
Fresh news