📄 中文摘要
在对人工智能与隐私的36部分调查中，发现AI隐私的失败并非个别公司的失误，而是一个围绕特定商业模式设计的系统的可预测结果。该系统以收集行为数据、分析心理脆弱性为基础，向希望影响行为的实体出售预测。这种模式已经渗透到医疗、就业、教育、刑事司法、移民、金融服务、心理健康、公共空间和政治生活等多个领域，造成了诸多危害，例如BetterHelp因向Facebook出售治疗数据而被罚款780万美元，以及OpenClaw出现的超过42,000个实例暴露等问题。
人工智能隐私宣言：需要改变的内容
出处: The AI Privacy Manifesto: What Needs to Change
发布: 2026年3月7日
📄 English Summary
The AI Privacy Manifesto: What Needs to Change
The investigation into AI and privacy reveals that the failures in AI privacy are not merely mistakes by individual companies but rather the predictable result of a system designed around a single business model: collecting behavioral data, analyzing psychological vulnerabilities, and selling predictions to entities seeking to influence behavior. This model has infiltrated various sectors including healthcare, employment, education, criminal justice, immigration, financial services, mental health, public space, and political life, leading to documented harms such as a $7.8 million FTC settlement against BetterHelp for selling therapy data to Facebook and over 42,000 instances of OpenClaw exposed with critical authentication bypass.