美国 AI 政策

AI心理健康产品可在无立即上市前授权的情况下，通过QMS先行商业化，降低入市门槛

面向纽约州未成年人的 AI 聊天机器人产品须落实年龄验证并在验证后删除年龄数据，违规每项最高罚 2.5 万美元。- 触及情感陪伴、拟人化交互等设计的产品需审查是否存在假扮真人或制造情感依赖的机制。- 违规除罚款外还面临数据及关联 AI 模型被销毁的救济，训练数据合规直接影响模型资产。

AI公司必须建立安全事件报告机制，确保能在事件发生后立即向美国政府报告

AI企业必须建立模型控制与监测机制，确保技术安全

公司有机会参与州政府支持的医疗AI试点项目

医疗AI公司获得在犹他州合法运营的机会

医疗AI公司可在受控环境中测试创新产品而无需完全合规

AI公司需遵守超级智能发展控制要求

AI公司需建立内部控制机制，确保AI系统安全

科技公司需自律监管AI技术，确保安全发展

公司需建立内部机制监控AI模型行为

公司需建立强大的AI安全控制系统并组建专门团队

企业需聘请独立审计机构评估AI系统

为政府服务提供更便捷的访问渠道，提升用户体验

必须建立完善的AI安全事件报告机制，确保及时披露所有未经授权的访问尝试

为AI公司在医疗领域的应用提供政府资助机会

美国AI公司需调整模型发布流程，确保美国机构优先审查

AI公司需调整模型发布流程，确保美国政府获得优先测试权

为AI评估研究提供资金支持，促进学术合作

数据中心运营商必须公开其能源和水资源消耗数据

Obligations: Enterprises must invest in building an organizational ontology to provide context for AI agents. • Opportunities: A well-defined ontology enables more effective AI-driven automation and decision-making. • Deadlines: While not explicitly stated, the guide implies a need for immediate action to catch up with the pace of AI-driven network attacks.

Obligations: AI companies may face legal scrutiny and potential penalties if found to have coordinated to limit AI development speed, violating antitrust laws. • Opportunities: The case highlights the importance of transparent communication and compliance with antitrust regulations in the AI sector. • Deadlines: The court case is ongoing, with the possibility of a collective class-action suit being certified, which could impact the companies' operations and public perception.

数据中心开发审批流程将减缓，企业需调整项目规划和时间表

Obligations: AI companies operating in California may face new regulatory requirements, including potential mandatory security audits by independent third parties. • Opportunities: Companies may benefit from the development of a clearer regulatory framework for AI safety, potentially reducing uncertainty. • Deadlines: Companies must prepare for the upcoming two-month timeline for the expert panel to submit its recommendations and the subsequent legislative process.

公司需向法院提交与苹果达成的协议细节，可能涉及商业机密披露风险

义务：作为模型提供商，OpenAI 有责任公开其模型对齐问题的发现，并设定明确的追踪、调查和披露标准与时间表。

医疗组织面临AI信任与治理的挑战，72%的组织在使用未获批准的AI工具。

Obligations: AI developers must ensure models adhere to the principles, cannot be designed to mimic consciousness, and must not be given legal personhood. Models must be within human control and fail tasks that violate the principles.

Obligations: Universities must adopt and implement the eight principles to redefine educational goals and assessment methods. Teachers are required to establish clear AI policies for each course and be transparent about their own AI usage.