Moonshot serves Claude instead of Kimi and collects exchanges for model training

The claim is that Moonshot allegedly routed users to Claude while presenting it as its own Kimi model, harvesting those exchanges for training — an accusation that, if accurate, raises serious questions about model provenance, terms-of-service compliance, and what customers are actually paying for. The accompanying commentary makes a broader point: such misuse only came to light because Anthropic investigated and published its findings. That dynamic matters for how the industry handles transparency going forward.

• Allegation that a rival AI provider served Claude under its own branding and collected the resulting conversations for training.
• Raises provenance and trust issues: users may not know which model is actually answering them.
• Disclosure of abuse depends on vendors voluntarily investigating and publishing; punitive coverage risks discouraging future transparency.