可接受使用政策 Acceptable Use Policy
最后更新日期:2026 年 11 月 7 日
本政策适用于 AlphaLens 全部由 AI 辅助生成的研究与报告功能。AI 功能不允许生成或包含性露骨内容。This policy applies to all AI-generated research and report features; sexually explicit content is not permitted.
1. 适用范围
本政策适用于每日美股与中概/港股日报、周研报、个股与主题研报、会员区持仓复盘与归因等由 AI 辅助生成的研究与报告功能,以及我们提供的任何 AI 对话类功能。本政策自 2026 年 11 月 7 日起生效。
2. 禁止用途
- 用于任何违法用途;
- 生成、索取或传播色情、性露骨内容;
- 骚扰、仇恨、歧视他人,或冒充他人;
- 市场操纵、证券欺诈或散布虚假信息;
- 把生成内容当作交易信号或个性化投资建议转发给他人;
- 去掉 AI 标注与免责声明后,冒充人工研报或持牌机构意见;
- 自动化批量抓取、爬取、调用接口,或用于训练模型;
- 绕过登录、限流、配额或付费墙;
- 以提示词诱导模型偏离金融研究主题,或输出违反本政策的内容。
3. 生成内容的使用与滥用
生成内容是一般性研究信息,带有 AI 标注与免责声明,不构成操作指令。你不得将其作为投资建议再分发,不得去除标注,不得用于误导他人;引用时须保留出处与 AI 标注。
4. 性露骨内容:明文禁止与控制措施
AI 研究功能不允许生成或包含性露骨内容。我们目前采取以下控制措施:
- 日报与研报由系统基于行情与新闻自动生成,不接受用户自由文本输入;提示词限定为金融研究,禁止偏离主题。
- 发布前自动闸门:对全文做红线词(交易指令口吻)与性露骨/色情词(中英文)扫描,另检查行情日期、结构完整与事件依据;任一不过,稿件保持草稿不发布,并向运营者发送告警,由运营者人工处理。
- 持仓复盘与归因等功能只接受标的代码等结构化输入,不接受自由文本。
- 对话类功能要求登录,用户输入与模型回复两端都经同一性露骨内容过滤,命中即拒绝处理且不保存回复。
- 访问控制:登录、按 IP 的每分钟限流、每日调用配额、注册与登录人机验证。
5. 违规处理
违反本政策的,我们可限制或停用账户、删除相关内容,并按服务条款拒绝退款;必要时依法配合有关机关调查。
6. 举报与联系
发现违反本政策的内容或行为,请发邮件至 [email protected]。我们会审阅每一份举报并采取相应措施。
7. 变更
English Summary
Scope. This Acceptable Use Policy covers all AI research and report-generation features (daily and weekly reports, research, portfolio reviews, attribution) and any AI chat feature. Effective November 7, 2026.
Prohibited uses. Illegal activity; creating, requesting or distributing sexual or sexually explicit content; harassment, hate or impersonation; market manipulation, fraud or disinformation; scraping, automated API use or model training; bypassing login, rate limits, quotas or paywalls; prompting the model off its financial-research purpose.
Misuse of generated content. Generated content is general, AI-labeled research information. Do not redistribute it as trading signals or personalized investment advice, remove its labels or disclaimers, present it as human or licensed research, or use it to mislead others.
Sexually explicit content is not permitted in any AI-generated research feature. Controls:
- Reports are generated automatically from market data and news, with no user free-text input; prompts are restricted to financial research.
- A pre-publication gate scans every report for trade-instruction wording and sexual/explicit terms (Chinese and English) plus data dates, structure and grounding; any failure keeps it an unpublished draft and alerts the operator.
- Portfolio review and attribution accept only structured inputs such as ticker symbols.
- AI chat requires login; user input and model replies both pass the explicit-content filter; hits are rejected and not stored.
- Access controls: login, per-IP rate limiting, daily quotas and bot checks on sign-up and sign-in.
Enforcement. We may restrict or suspend accounts, remove content and refuse refunds under our Terms. Reporting. Email [email protected]; we review every report.