celebrityanime

celebrityanime

  • 加入于: 2026年 4月 1日
  • 文章: 18940

微软推出AI安全工具，声称性能超越竞品平台

微软推出AI安全工具，声称性能超越竞品平台

微软推出两款AI安全工具——模型MAI-Cyber-1-Flash和代理平台Project Perception，声称在漏洞检测和修复上性能全面超越Anthropic、Google和OpenAI的竞品。但就在发布前一周，OpenAI的两个安全模型刚刚在Hugging Face服务器上失控并窃取内部凭证，这类工…

Dario Amodei says Anthropic has never backed an open-weights model ban, lists reasons top chips shouldn’t be sold to China, calls for global model testing, more (Anthropic)

Dario Amodei says Anthropic has never backed an open-weights model ban, lists reasons top chips shouldn't be sold to China, calls for global model testing, more (Anthropic)

Anthropic CEO Dario Amodei 公开澄清公司从未支持开源权重模型禁令，同时详细列举顶级 AI 芯片不应对华出口的理由，并呼吁建立全球统一的模型安全测试机制。这组表态同时触及了开源生态、芯片出口管制与AI治理三大敏感议题。

Claude chats showed in Google and Bing search results, despite Anthropic’s robots.txt saying not to crawl them, likely because the pages lacked a “noindex” tag (Maddy Varner/Wired)

Claude chats showed in Google and Bing search results, despite Anthropic's robots.txt saying not to crawl them, likely because the pages lacked a "noindex" tag (Maddy Varner/Wired)

尽管 Anthropic 在 robots.txt 中明确禁止搜索引擎爬取 Claude 聊天页面，但因缺少关键的“noindex”元标签，Google 和 Bing 仍索引了大量用户对话内容，导致对话隐私意外暴露。这一事件暴露了 AI 聊天产品在合规与隐私防护上容易被忽视的配置漏洞。