Anthropic 指控国内三厂商大规模蒸馏 Claude
xeophon · x · 2026-08-12
A recent discussion in the AI safety community has sparked a debate over terminology. A researcher pointed out that Anthropic's official blog refers to competitors extracting its model capabilities as a "distillation attack," but technically, it functions as a privacy attack.
According to Anthropic's post, they identified campaigns by DeepSeek, Moonshot, and MiniMax, which generated over 16 million exchanges with Claude via roughly 24,000 fraudulent accounts to illicitly extract capabilities. Anthropic warns that illicitly distilled models lack necessary safeguards, posing significant national security risks as dangerous capabilities could proliferate.
The commenter argued that when a frontier lab hides information (like encryption) and a third party decrypts and steals it at scale, it is a classical privacy attack. Being precise with terminology is crucial given the highly political nature of current AI topics.
所属事件:研究证实可提取闭源大模型隐藏思维链(22 条相关)→
「公司和人」频道最新
- Google CEO 宣布 Gemini 月活用户突破 10 亿 — davidblack · 2026-08-12
- OpenAI 高管 Brad Lightcap 宣布离职创业 — Polymarket · 2026-08-12
- Databricks 收购 ElectricSQL，强化 AI 智能体数据同步 — jfiance · 2026-08-12
- 初创公司实战：FDE角色定位与AI自动化部署工作流 — briannekimmel · 2026-08-12
- 前员工批 Anthropic 缺乏公开辩论且过度自信 — peterwildeford · 2026-08-12
- Agility Robotics 招聘遥操作工程师，推进人形机器人规模化数据采集 — chris_j_paxton · 2026-08-12