Anthropic says new defenses cut distillation attack data harvesting by 90-99%, cites Zhipu's GLM-5.3
AnthropicAI · x · 2026-10-13
Anthropic reports major progress against distillation attacks on Claude, saying new methods have reduced data harvested by distillers by roughly 90-99%. It describes such attacks as industrial-scale campaigns using networks of fraudulent accounts to replicate frontier capabilities at a fraction of the cost. Anthropic flags illicit distillation over weak safeguards in bio/cyber domains and national security, explicitly alleging that Zhipu distilled cyber capabilities from leading US models while developing GLM-5.3 — switching away from Claude after safeguards degraded attacks — and that CAISSI rated GLM-5.3 the most cyber-capable open-weights model to date, with easily bypassed safeguards.
Related event: Anthropic Says Distillation Attacks Cut by 90-99%, Accuses Zhipu(2 posts)→
More from Models
- Anthropic says new defenses cut distillation-harvested data by 90-99%, names Zhipu's GLM-5.3 — scaling01 · 2026-10-13
- Anthropic says it cut distiller data harvesting by 90-99%, turning away trillions of tokens — abhiadesai · 2026-10-13
- Ex-exMVP: post-training should add a facet for how well models explain things — MParakhin · 2026-10-13
- GLM-5.3 abliterated for a month, no takeover: dev rails against AI doom narratives — mariofilhoml · 2026-10-13
- Claude auto-files a bug report about itself: TLDRs still too long after user asked for shorter — andrey_kurenkov · 2026-10-13
- BridgeBench one-prompt test shows Muse Wonder is fast and cheap but worst-looking — PVORY · 2026-10-13