OpenAI Blocked 15,000 Accounts Stealing Model Reasoning, but the Attack Worked on Azure for Weeks
The Decoder · rss · 2026-10-01
OpenAI says it stopped a coordinated campaign in which 15,000+ accounts tried to extract the hidden reasoning of its models, tying part of the activity to people connected to Moonshot AI.
However, researchers found the same attack kept working on Microsoft Azure for weeks — even against the new GPT-6 Astra — exposing a gap: OpenAI's protections don't extend to cloud platforms also selling its models.
More from Safety
- Massive supply-chain attack hits Bitrise, OpenCode, Railway, Northflank, Fireworks hunting AI keys — Afinetheorem · 2026-11-16
- Reprogramming fewer than 10 dormant neurons backdoors any LLM agent, no training needed — DimitrisPapail · 2026-11-16
- 'Structure transfer': using an NTSB report scaffold to evade Pangram AI detection — amplifiedamp · 2026-11-16
- Open-source Writing Scaffold MCP library helps AI writing bypass Pangram detector, up to 96% effective — amplifiedamp · 2026-11-16
- Pangram AI-text detector bypass called trivial-but-costly; false-positive search also feasible — ctjlewis · 2026-11-16
- HN job seeker gets RAT and info stealer from a fake recruiter — GetJump · 2026-11-16