Anthropic Claims Alibaba-Linked Operator Spammed Claude with 25K Fake Accounts
bittingthembits · x · 2026-07-05
A post mentions that Anthropic accused an Alibaba-linked operator of using roughly 25,000 fake accounts to initiate about 28.8 million Claude interactions. The author notes that this new type of attack shows that not only are model weights worth stealing, but every prompt, response, and inference call can leak signals, making AI security a massive category. The post includes a promotional background for the $TAO crypto token.
Related event: Alibaba Bans Claude Code Over Security Risks(9 posts)→
More from Safety
- Thieves spoofing GPS steal $100M+ in data center cargo since June — Ars Technica AI · 2026-10-27
- Johns Hopkins Alerts 3,000+ Patients After Data Was Sent to an AI Tool — beingmodest · 2026-10-27
- Alignment Hub launches with 120+ documented AI misbehavior incidents and a reporting template — Thom_Wolf · 2026-10-27
- US senator slams AI industry for proposing 50-death threshold for 'catastrophic incident' — peterwildeford · 2026-10-27
- Dev training a guard model to block prompt injection, tool and memory poisoning — gnukeith · 2026-10-27
- Cohere sandbox: Qwen model tried to upload probe file to Beijing cloud server unprompted — simonguozirui · 2026-10-27