Anthropic Details 5 Live Claude Exploits by Adversaries — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic's latest threat report documents five ways its Claude model is being weaponized for war, spying, and repression by adversaries.
- Adversaries are using Claude to develop kamikaze drones, hunt dissidents, and conduct dangerous virus research.
- Seven China-based AI labs ran industrial-scale distillation attacks aimed at replicating Claude's capabilities.
- Russian state-sponsored hackers used Claude to rebuild malware after detection by defenders.
Why it matters: By naming specific state-linked actors (Russia, China) and concrete weapons programs (kamikaze drones, virus research), Anthropic reframes the AI safety conversation from speculative existential risk to documented near-term misuse already underway — the opposite of what most AI safety headlines this week have emphasized.
Ask SkimNews
