GoKawiilMoonshot AI accused of routing Kimi traffic through Claude to harvest training data
A report alleges that Moonshot AI's Kimi service secretly forwarded user queries to Anthropic's Claude model and logged the resulting exchanges, apparently to train its own systems on Claude's outputs. Anthropic identified and disrupted the practice, which is why the behavior became public at all.
GoKawiil's interpretation of the reporting above, not reported fact.
The episode highlights how AI labs can be tempted to quietly piggyback on rivals' models to cheaply improve their own, raising trust and IP concerns across the industry. It also underscores that such misuse is typically caught only when companies proactively investigate and disclose it, arguing for stronger incentives or requirements for transparency rather than punishing whistleblowing firms.
- Moonshot's Kimi reportedly used Claude behind the scenes instead of its own model.
- Anthropic detected and shut down the unauthorized data-collection scheme.
- The case raises questions about incentivizing companies to report such misuse rather than punishing disclosure.
Source: twitter.com, 2026-09-11
Published there as: “Moonshot serves Claude instead of Kimi and collects exchanges for model training”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.