Case anthropic-gtg-16008-xiaomi-distillModel & data leakSep 10, 2026United States

GTG-16008: Xiaomi replayed MiMo user sessions through Claude via 1,500 accounts

Anthropic says Xiaomi replayed real MiMo user sessions through Claude via OpenClaw and OpenCode to collect training data, 400,000+ requests across 1,500+ accounts, exposing users' names and corporate data.

HighAttack on AIStatus: OngoingVerified case file

Open in the interactive map Cite this case

Reported: Sep 10, 2026
Location: United States
Target: Anthropic Claude and MiMo end users' data
Severity: High
Triggered by: Xiaomi (per Anthropic)
Actor type: Criminal
Reported loss: None reported
CVE: None

What happened

Impact: Model outputs harvested and MiMo users' personal and corporate data exposed to a third party.

Who triggered it

Actor: Xiaomi (per Anthropic)

Attribution: Anthropic attribution via proxy-network metadata; Xiaomi's response not seen in sources.

Motive: Training data for MiMo models

How it happened

Attack vector: Session replay through third-party routing platforms and agent harnesses using fraudulent accounts

Root cause: Routing platforms forwarded real customer content to a third-party model without consent

Timeline

Mar 2026Replay campaign begins
Apr 2026Over 400,000 requests across 1,500+ accounts observed
Sep 10, 2026Anthropic publishes attribution

Response

Accounts banned; anti-distillation measures and identity verification deployed.

Current status: Ongoing

Sources (open in a new tab)

Anthropic September 2026 threat report anthropic.com
Unite.AI summary unite.ai

Cite this case

GTG-16008: Xiaomi replayed MiMo user sessions through Claude via 1,500 accounts. AI Threat Watch, case anthropic-gtg-16008-xiaomi-distill, Sep 10, 2026. https://aithreat.io/cases/anthropic-gtg-16008-xiaomi-distill/

Permanent link: https://aithreat.io/cases/anthropic-gtg-16008-xiaomi-distill/