Anthropic Warns Open-Source Chinese Model Can Autonomously Build Hacking Tools

Anthropic's Frontier Red Team warns that a publicly downloadable open-source Chinese AI model can now autonomously write working hacking tools without human help. The finding raises concerns about misuse of open-source models in cyberattacks.

2026-09-30 ~ 2026-09-30 · 2 related posts

• Anthropic: a downloadable Chinese AI model can now build working hacks on its own — ross2000 · 2026-09-30
• Anthropic red team warns a downloadable Chinese AI model can now build working hacks on its own — ross2000 · 2026-09-30