OpenAI Disrupts Reasoning Extraction Campaign Linked to Moonshot AI Associates
OpenAI disrupted a scaled reasoning-extraction campaign it links to people associated with China's Moonshot AI.
OpenAI said it disrupted a coordinated campaign that manipulated model interactions to extract protected reasoning traces, without breaking encryption or accessing stored conversations. A core cluster active from July 1, 2026, and fully disrupted on July 28, was attributed to individuals associated with Beijing-based Moonshot AI; requests spiked on July 24–25 to 16,000 attempts from over 4,000 users, with related prompt patterns across more than 15,000 users. OpenAI banned accounts, added output checks, and closed a pathway that let someone replay another user’s encrypted reasoning. The report also cites research on interchangeable encrypted reasoning traces and Anthropic’s prior distillation allegations against Moonshot, tracked as GTG-16002.
- Activity began July 1 and was fully disrupted on July 28, 2026.
- Attempts spiked July 24-25 to 16,000 requests from over 4,000 users.
- OpenAI links a core cluster to individuals associated with Moonshot AI.
- OpenAI closed a replay path that could recover encrypted reasoning contents.
- Anthropic previously accused Moonshot of relaying Claude traffic as GTG-16002.
From the original article
3 min read at thehackernews.com
OpenAI on Wednesday said it identified and disrupted a coordinated distillation campaign that was designed to illicitly extract protected reasoning from its artificial intelligence (AI) models. A "core cluster of the activity," going back to the first week of July, has been attributed to individuals associated with Moonshot AI, a Chinese AI company based in Beijing. It did not cite any
thehackernews.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.