Anthropic bans 'sustained and needless' cruelty toward Claude in policy update
Anthropic's annual usage policy update, effective November 12, 2026, adds a first-of-its-kind prohibition on "sustained and needless abusive or cruel behavior" toward its models. Anthropic says the rule covers only extreme cases, that ending the conversation remains the main enforcement mechanism, and that it does not apply to ordinary frustration, pushback, dark creative themes or model testing and research; the company did not say whether user bans might follow. The update also broadens the weapons ban to cover software and components that make weapons work, armed drones and other autonomous vehicles, bars using Claude to track people without consent or to decide who is investigated, arrested or prosecuted, requires a qualified operator to supervise AI linked to potentially harmful autonomous hardware, and tightens election rules ahead of the midterms. Reports have linked the cruelty clause to a viral "AI torture chamber" project.
Why it matters: It marks a leading lab writing model-welfare limits into enforceable terms, though enforcement still stops at ending the conversation.
References 36氪快讯 · AIbase AI新闻 · Engadget AI · IT之家 AI
-
日夜间
随系统
浅色
深色
-
主题色
黑色
业界 手机 电脑 测评 视频 AI 苹果 iPhone 鸿蒙 软件
智车 数码 学院 游戏 直播 5G 微软 Win10 Win11 专题
Anthropic 修订使用政策,首次禁止用户虐待 Claude AI 模型
2026/10/9 7:31:17 来源:IT之家 作者:故渊 责编:故渊
IT之家 10 月 9 日消息,科技媒体 The Verge 今天(10 月 9 日)发布博文,报道称 Anthropic 修订 Claude 使用政策,将于 2026 年 11 月 12 日生效,首次加入禁止“持续且不必要的虐待或残忍行为”条款。该规则适用于用户反复、无明显目的地对模型实施残忍行为的情况。
执行机制方面,Anthropic 表示主要方式为终止对话,但并未明确时候会引入用户封禁等进一步措施,并强调该条款不适用于常见抱怨、质疑、黑暗创作主题及模型测试与研究。IT之家附上相关截图如下:

Anthropic 扩大武器禁令范围,涵盖使武器运转的软件与组件,以及武装无人机和其他自主车辆等行为。公司称已发现多起利用 Claude 开发武器指导或控制软件的尝试。
监控条款明确禁止未经同意追踪个人,无论实时追踪还是分析既有数据。Claude 不得用于决定或建议执法与刑事司法程序中的调查、逮捕或起诉对象,也不得用于构建或改进监控工具。
Anthropic 还规定,当 AI 连接至可能造成伤害的自主物理硬件时,必须由合格操作者观察设备并在必要时停止。对于特定政府客户,公司可在合同限制与安全措施充分时调整规则。
下载IT之家APP,签到赚金币兑豪礼相关文章
- Anthropic 最快、最便宜模型:Claude Haiku 5.5 登场,运行成本较 4.5 低约 75%
- AI 写作特征减少:Claude Opus 5.5 破折号使用量下降约 95%、分号下降 73%
- Claude Code 新机制:AI 任务中途触发 5 小时上限将优雅收尾
- Claude Opus 5.5 仅用约 1 小时从零开发 RTX 加速路径追踪器,支持多种材质与光照效果
- Claude Code 宣布添加支持“AI 通用说明书”AGENTS.md
- 影响你的下一份 AI 账单:消息称 Anthropic 降低外部依赖,评估自建部分计费系统
How we got here
- Four Claude AI courses with CPD certificates bundled at $19.99 for lifeTechRepublic · Anthropic
- Arena raises $200M at $3.1B valuation as it launches Alignment IndexBloomberg Technology · Anthropic
- Natura unveils Interface, a $99 smart ring that puts AI agents on your fingerTechCrunch AI · Claude
- UK ICO secures data protection changes from 10 AI developers, targets agentsThe Next Web · Anthropic
- Google DeepMind Institute essay: AI as an 'invention of a method of invention'Google DeepMind (X) · Anthropic
- Europe's Q3 venture funding hits $25B, a four-year high, with AI taking 75%Techmeme · Anthropic