美国纽约州正式启动《RAISE法案》的实施工作,标志着对前沿人工智能开发企业的监管进入新阶段 1。该法案明确要求年收入超过5亿美元的大型前沿AI开发商(包括OpenAI和Anthropic)必须在2027年1月1日前通过指定门户完成注册 1。作为监管执行机构,新成立的数字创新、治理、诚信与信任办公室(DIGIT)将负责监督合规情况,单次违规罚款上限为100万美元,后续违规最高可达300万美元 1。
根据法案规定,受监管企业需公开其安全框架,在发生重大安全事件后72小时内进行报告,并发布模型透明度报告 1。为支持独立评估先进AI系统的安全性,位于宾汉姆顿大学的“纽约AI责任与研究中心”同步启动,初始资金包含5500万美元的捐赠和资本 1。这一举措使得纽约州与其他地区形成差异化的监管格局:加州SB 53法案已于今年1月生效,规定重大安全事件报告期限为15天;而伊利诺伊州SB 315法案计划于2027年初生效 1。此外,纽约州此前颁布的针对AI伴侣及未成年人聊天机器人的其他安全法律将于2027年6月1日生效 1。
New York State has begun implementing the RAISE Act, a law requiring large frontier AI developers to register and disclose safety measures. 1 The legislation targets companies with annual revenues exceeding $500 million, such as OpenAI and Anthropic, which must complete registration via a portal by January 1, 2027. 1 Registered firms are mandated to publish their safety frameworks, issue model transparency reports, and report major security incidents within 72 hours. 1
Enforcement is handled by the newly created Office of Digital Innovation, Governance, Integrity, and Trust (DIGIT). 1 Violations carry fines of up to $1 million per incident, rising to a maximum of $3 million for subsequent offenses. 1 To independently evaluate advanced AI system safety, New York also launched the New York AI Accountability and Research Center at Binghamton University, supported by an initial $55 million in donations and capital. 1
The RAISE Act joins other state-level regulations; California’s SB 53 took effect in January, requiring major security event reporting within 15 days, while Illinois’ SB 315 will become effective in early 2027. 1 Additionally, New York has enacted separate laws regarding AI companions and chatbots for minors, with the latter scheduled to take effect on June 1, 2027. 1
评论
还没有评论,欢迎留下第一条。