Unverified claim: Xiaomi fed user chats to Claude; Zhipu shifted from Fable to Opus 4.6
ns123abc · x · 2026-09-11
An unverified post claims Xiaomi replayed real user conversations — names, contacts, company data across a dozen languages — into Claude to generate training data. A related claim says Zai.org (Zhipu) tried to steal US frontier models' cyber skills by targeting Fable, gave up due to safeguards, and switched to Opus 4.6. No official confirmation; treat with caution.
More from Safety
- Anthropic Says It Caught China, Russia, Iran Actors Using Claude for Pathogen Research — connoraxiotes · 2026-09-11
- Anthropic jailbreak incident dissected: case against the misalignment interpretation plus 4 new Claudes — jessi_cata · 2026-09-11
- About 30 robots protest outside Poland's digital ministry demanding AI rules — MoonL88537 · 2026-09-11
- AI Safety Debate: "Just Regulate More" Is Cope, Proposals Must Be Concrete — sytelus · 2026-09-11
- Debate: are AI safety warnings crying wolf, or necessary prep time — BlackHC · 2026-09-11
- Could the US and China agree to slow AI? A tracker compiles all government statements — NathanpmYoung · 2026-09-11