OpenAI’s Astra reported as a major target for Chinese AI labs
AI · First reported 29 Oct, 01:00 · Updated 17:49
The Decoder reports that DeepSeek and Moonshot AI relied heavily on OpenAI’s GPT-6 Astra outputs to train models, while Anthropic says its defenses have reduced the data attackers can obtain from Claude. OpenAI has also reported a distillation campaign it says involved people linked to Moonshot.
- The Decoder reports that DeepSeek and Moonshot AI used Astra outputs extensively to train models including Kimi K3.1 Pro and V4.1-Flash 1017.
- Anthropic says its newer defenses reduced the data attackers could obtain from Claude by 90 to 99 percent.
- OpenAI said in late September that it had identified a coordinated distillation campaign involving people it linked to Moonshot; Moonshot declined to comment and DeepSeek did not respond to a request for comment.
Summary written by Lux AI from 2 articles
Key facts
- Models identified as Astra training targets
- Kimi K3.1 Pro and V4.1-Flash 1017
- Anthropic’s stated reduction in data attackers can obtain
- 90 to 99 percent
- OpenAI’s reported campaign announcement
- Late September; OpenAI said it involved people linked to Moonshot
Source spectrum
Far left
Left
Lean left
Centre
Lean right
Right
Far right
Left 0 · Centre 2 · Right 0 of 2 rated sources
All coverage
Coverage over time
LeftCentreRight
Articles per hour since first reported, by side.
Ownership
- Corporation 100% (2)
Share of the 2 sources by owner type.